Professional Documents
Culture Documents
NATO's Air War For Kosovo - A Strategic and Operational Assessment-Rand Publishing (2001)
NATO's Air War For Kosovo - A Strategic and Operational Assessment-Rand Publishing (2001)
Nis
l
Raska l l
Blace
l
Sjenica K
L
O
P
Prokuplje
A
O
l
l
Zitorada
N
IK Doljevac
l
l Brestovac l
S
l
Kursumlija
Novi
A
Pazar
N
S e r b i a Bojnik
D
l
l
Leskovac
Zˇ
l
Tutin
A
Lebane l
Montenegro l Kosovska
Mitrovica Podujevo
l
K
Istok l K
O
Srbica l S
l
l
Poljance O G
l Djurakovac V O
Pec Vitomirica O L J
l Pristina A K
Klina Glogovac Kosovo l Novo
l
l Poljec Brdo Kosovska
M E
l
l Gracanica l Kamenica
l
Plav l
P
l
Lapusnik
l
Janjevo Vranje l
T
Decani l Lipljan
O
o
L
N
.A
J
H I
LB
Malisevo l
E
l
AN
Junik Gnjilane
l l
I
AN
Bujanovac
J A
AL
PS
l
Bajram Orahovac
l
Curril Djakovica l
l Presevo
Vitina l
Kosovo
Kacanik
l
A L B A N I A
CR
l Prizren
NA
Shalqin l
l
Blace Kumanovo l
GO
Kukes
RA
l
Dragas A
IN
l
N
A Tetovo l
L
P
l
Brod Skopje
R
l
Kamenjane
SA
Gostivar l
MACEDONIA
l
Titov
Veles
Map of Kosovo
PREFACE
v
vi NATO’s Air War for Kosovo: A Strategic and Operational Assessment
mented herein was carried out in Project AIR FORCE’s Strategy and
Doctrine Program and was completed in August 2001. All
photographs included in this study were provided by the U.S.
Department of Defense. The book should be of interest to USAF
officers and other members of the U.S. national security community
concerned with strategy and force employment issues raised by
NATO’s air war for Kosovo and with the implications of that
experience for force development, air power doctrine, and concepts
of operations for joint and coalition warfare.
xi
SUMMARY
Between March 24 and June 9, 1999, NATO, led by the United States,
conducted an air war against Yugoslavia in an effort to halt and re-
verse the human-rights abuses that were being committed against
the citizens of its Kosovo province by Yugoslavia’s president, Slobo-
dan Milosevic. That 78-day air war, called Operation Allied Force,
represented the third time during the 1990s in which air power
proved pivotal in determining the outcome of a regional conflict. Yet
notwithstanding its ultimate success, what began as a hopeful
gambit for producing Milosevic’s quick compliance soon devolved,
for a time at least, into a seemingly ineffectual bombing experiment
with no clear end in sight. Not only was the operation’s execution
hampered by uncooperative weather and a surprisingly resilient op-
ponent, it was further afflicted by persistent hesitancy on the part of
U.S. and NATO political leaders and sharp differences of opinion
within the most senior U.S. military command element over the most
effective way of applying allied air power against Serb assets. More-
over, the plan ultimately adopted ruled out any backstopping by al-
lied ground troops because of concerns over the potential for a land
invasion to generate unacceptable casualties and the consequent low
likelihood of mustering the needed congressional and allied support
for such an option. All planning further assumed that NATO’s most
crucial vulnerable area was its continued cohesion. Therefore, any
target or attack tactic deemed even remotely likely to undermine that
cohesion, such as the loss of friendly aircrews, excessive collateral
damage, or anything else that might weaken domestic support, was
to be most carefully considered, if not avoided altogether. All of that,
however unavoidable some aspects of it may have been, made
xiii
xiv NATO’s Air War for Kosovo: A Strategic and Operational Assessment
NATO’s air war for Kosovo a step backward in efficiency when com-
pared to the Desert Storm air campaign.
One can, of course, insist that air power alone was the cause of Milo-
sevic’s capitulation in the tautological sense that Allied Force was an
air-only operation and that in its absence, there would have been no
reason for believing that he would have acceded to NATO’s de-
mands. Yet as crucial as the 78-day bombing effort was in bringing
Milosevic to heel, one should be wary of any intimation that NATO’s
use of air power produced a successful result for the alliance without
any significant contribution by other factors. For example, beyond
the obvious damage that was being caused by NATO’s air attacks and
the equally obvious fact that NATO could have continued bombing
both indefinitely and with virtual impunity, another likely factor be-
hind Milosevic’s capitulation was the fact that the sheer depravity of
Serbian conduct in Kosovo had stripped the Yugoslavian leader of
any remaining vestige of international support, including, in the end,
from his principal backers in Moscow.
On top of that was the sense of walls closing in that Milosevic must
have had when he was indicted as a war criminal by a UN tribunal
only a week before his loss of Moscow’s support. Yet a third factor
may have been the mounting pressure from Milosevic’s cronies
among the Yugoslav civilian oligarchy, prompted by the continued
bombing of military-related industries, utilities, and other infrastruc-
ture targets in and around Belgrade in which they had an economic
stake and whose destruction increasingly threatened to bankrupt
them.
Finally, one must take into account what Milosevic no doubt per-
ceived, rightly or wrongly, to have been the possibility of an eventual
Summary xv
Some, however, have made more of that fact than the evidence war-
rants. In the early wake of the successful conclusion of Allied Force,
revisionist claims began emanating from some quarters suggesting
that the air effort had been totally ineffective and that, in the end, it
had been Milosevic’s fear of a NATO ground invasion that had in-
duced him to capitulate. Those claims defy believability because any
NATO ground invasion, however probable it may have been in the
end, would have taken months, at a minimum, to prepare for and
successfully mount.
Senior civilian defense officials and U.S. Air Force leaders freely con-
ceded after the Serbian withdrawal that the problems encountered
by the largely failed effort against fielded enemy forces reflected real
challenges for the effective application of air power posed by such
impediments as trees, mountains, poor weather, and an enemy
ground force that is permitted the luxury of dispersing and hiding
rather than concentrating to maneuver to accomplish its mission.
Yet while it was essential for NATO to try its best to keep Serb forces
pinned down and incapable of operating at will, the majority of the
sorties devoted to finding and attacking enemy troops in Kosovo en-
tailed an inefficient and ineffective use of munitions and other valu-
able assets. That said, the targeting of enemy ground forces operat-
ing within Kosovo was an inescapable political necessity, considering
that those forces were responsible for committing the ethnic cleans-
ing acts that NATO had vowed to stop. Failure to target those forces
would almost certainly have caused the bombing effort to lose cred-
ibility in the eyes of the NATO civilian leadership.
To begin with, the conduct of the air war as an allied effort came at
the cost of a flawed strategy that was further hobbled by the manifold
inefficiencies that were part and parcel of conducting combat opera-
tions by consensus. In addition to the natural friction created by
NATO’s approach to target approval, the initial reluctance of its
political leaders to countenance a more aggressive air campaign in
terms of target numbers and force size failed completely to capitalize
on air power’s potential for taking down entire systems of enemy
capability simultaneously. Further compounding the inefficiency of
Summary xix
Because NATO had initially hoped that the operation would last only
a few days, it failed to establish a smoothly running mechanism for
target development and review until late April. Once NATO’s going-
in assumption proved hollow, a frenetic rush ensued to come up
with additional target nominations that could be more quickly and
easily approved by NATO’s political authorities. Even then, there
was little by way of a consistently applied strategy behind the target
development process. Most of the attack planning done throughout
the air war was not driven by desired effects, but rather entailed
simply parceling out sortie and munitions allocations by target cate-
gory as individual targets were approved, without much considera-
tion given to how a target’s neutralization might contribute toward
advancing the overall objectives of the air war.
It was not only the alliance-induced friction that helped make for an
inefficient bombing effort. As Allied Force unfolded, it became in-
creasingly clear that even the U.S. military component was divided in
a high-level struggle over the most appropriate targeting strategy—
reminiscent of the feuding that had occurred nine years earlier be-
tween the Army’s corps commanders and the joint force air compo-
nent commander (JFACC), then–Lieutenant General Charles Horner,
over the ownership and control of air operations in Desert Storm.
Once the initial hope that Milosevic would fold within a few days af-
ter the bombing started proved groundless, NATO was forced into a
scramble to develop an alternative strategy. The immediate result
was an internecine battle between the Supreme Allied Commander
in Europe, U.S. Army General Wesley Clark, and his air component
commander, USAF Lieutenant General Michael Short, over where
the air attacks should be primarily directed. Short maintained that
the most effective use of allied air power would be to pay little heed
to dispersed Serbian forces in Kosovo and to concentrate instead on
infrastructure targets in and around Belgrade, including key electri-
cal power plants and government ministries. However, Clark in-
xx NATO’s Air War for Kosovo: A Strategic and Operational Assessment
achievement for air power. Yet with all due respect for the un-
matched professionalism of the aircrews who actually carried out the
air war, it is hard to accept that characterization as the proper con-
clusion to be drawn from Allied Force. To be sure, there is much to
be said of a positive nature about NATO’s air war for Kosovo. To be-
gin with, it did indeed represent the first time in which air power co-
erced an enemy leader to yield with no friendly land combat action
whatsoever. This does not mean that air power can now “win wars
alone” or that the air-only strategy ultimately adopted by NATO’s
leaders was the wisest choice available to them. Yet the fact that air
power prevailed on its own despite the multiple drawbacks of a re-
luctant administration, a divided Congress, an indifferent public, a
potentially fractious alliance, a determined opponent, and—not
least—the absence of a credible NATO strategy surely testified that
the air instrument has come a long way in recent years in its relative
combat leverage compared to that of other force elements in joint
warfare.
Although the manner in which the air war was conducted fell short of
the ideal use of air power, it suggested that gradualism may be here
to stay if U.S. leaders ever again intend to fight with coalition part-
ners for marginal or amorphous interests. Insofar as gradualism
promises to be the wave of the future, it suggests that airmen will
need to discipline their natural inclination to bridle whenever politi-
cians moderate the application of a doctrinally pure campaign strat-
egy and to recognize and accept instead that political considerations,
after all, determine—or should determine—the way in which cam-
paigns and wars are fought. This does not mean that military leaders
should surrender to political pressures without first making their
best case for using force in the most effective and cost-minimizing
way. It does, however, stand as an important reminder that war is ul-
timately about politics and that civilian control of the military is an
inherent part of the democratic tradition. Although air warfare pro-
fessionals, like all other warfighters, are duty-bound to try to per-
suade their civilian superiors of the merits of their recommenda-
tions, they also have a duty to live with the hands they are dealt and
to bend every effort to make the most of them in an imperfect world.
It follows that civilian leaders at the highest levels have an equal obli-
gation to try to stack the deck in such a way that the military has the
best possible hand to play and the fullest possible freedom to play it
to the best of its ability. This means expending the energy and politi-
Summary xxiii
To admit that gradualism may be the wave of the future for any U.S.
involvement in coalition warfare is hardly to accept that it is any
more justifiable from a military point of view for that reason alone.
Quite to the contrary, the incrementalism of NATO’s air war for
Kosovo, right up to its very end, involved a potential price that went
far beyond the loss of valuable aircraft and other expendables for
questionable gain. It risked frittering away the hard-earned reputa-
tion for effectiveness that U.S. air power had finally earned for itself
in Desert Storm after more than three years of unqualified misuse
over North Vietnam a generation earlier.
xxv
xxvi NATO’s Air War for Kosovo: A Strategic and Operational Assessment
xxix
xxx NATO’s Air War for Kosovo: A Strategic and Operational Assessment
TF Task Force
TIP Tactical Integrated Planning
TLAM Tomahawk Land-Attack Missile
TOT Time on Target
UAV Unmanned Aerial Vehicle
UCAV Unmanned Combat Air Vehicle
UHF Ultra-High Frequency
UN United Nations
USA United States Army
USAF United States Air Force
USAFE United States Air Forces in Europe
USAFE/SA United States Air Forces in Europe, Studies
and Analysis Office
USAREUR United States Army in Europe
USEUCOM United States European Command
USMC United States Marine Corps
USN United States Navy
USS United States Ship
VJ Yugoslav Army
VTC Video Teleconference
Chapter One
INTRODUCTION
Between March 24 and June 9, 1999, NATO, led by the United States,
conducted an air war against Yugoslavia in an effort to halt and re-
verse the continuing human-rights abuses that were being commit-
ted against the citizens of its Kosovo province (see the Frontispiece,
Map of Kosovo) by Yugoslavia’s elected president, Slobodan Milose-
vic. As it turned out, that 78-day effort, called Operation Allied Force,
represented the third time in a row during the 1990s, after Opera-
tions Desert Storm and Deliberate Force, in which air power proved
pivotal in determining the outcome of a regional conflict. Yet
notwithstanding its ultimate success, what began as a hopeful gambit
for producing quick compliance on Milosevic’s part soon devolved,
for a time at least, into a seemingly ineffectual bombing experiment
with no clear end in sight. Not only was the operation’s execution
hampered by uncooperative weather and a surprisingly resilient
opponent, it was further afflicted by persistent hesitancy on the part
of U.S. and NATO decisionmakers that was prompted by fears of
inadvertently killing civilians and losing friendly aircrews, as well as
by sharp differences of opinion within the most senior U.S.
command element over the best way of applying allied air power
against Serb assets to achieve the desired effects. All of that and
more, however unavoidable some aspects of it may have been, made
NATO’s air war for Kosovo a substantial step backward in efficiency
when compared to Desert Storm.
This book assesses Operation Allied Force from a strategic and op-
erational perspective, with a view toward spotlighting what was most
gratifying about the application of allied air power throughout the ef-
fort, as well as identifying and exploring aspects of air power’s per-
1
2 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Toward that end, the book first describes the air war’s strategic and
operational highlights in chronological order. It then considers the
various factors that interacted to induce Milosevic to capitulate when
he did. After that, it explores air power’s principal accomplishments,
as well as the many problems that worked to render Allied Force a
______________
1Christopher Cviic, “A Victory All the Same,” Survival, Summer 2000, p. 174.
Introduction 3
______________
2 In the latter respect, this assessment consciously seeks to avoid the common
syndrome of so-called lessons-learned efforts whereby “losers tend to study what went
wrong while winners study what went right.” Princeton University political scientist
Bernard Lewis, an adviser to the USAF’s Gulf War Air Power Survey conducted in
1991–1992, called this cautionary reminder to the attention of the survey team as that
effort was getting under way. Quoted in Gian P. Gentile, How Effective Is Strategic
Bombing? Lessons Learned from World War II to Kosovo, New York, New York
University Press, 2001, p. 182.
This page intentionally blank
Chapter Two
PRELUDE TO COMBAT
______________
1 For informed insight into the origins of the ancestral hatreds that animated the
atrocities committed against the Kosovar Albanians by Serbia, one can do no better
than the epic novel by Ivo Andric, The Bridge on the Drina, Chicago, University of
Chicago Press, 1977. Written in Serbo-Croatian in 1945, this tour de force won the
1961 Nobel prize for literature. It speaks about Balkan conflicts from the earliest
clashes between the Bosnian Turks and Serb Christians in the early 15th century to the
coming of the First World War. In a passage hauntingly reminiscent of more recent
Balkan horrors, Andric described how ethnic rivals as far back as the 17th century
“were as if drunk with bitterness, from desire for vengeance, and longed to punish and
kill whomsoever they could, since they could not punish or kill those whom they
wished” (pp. 86–87). Of a later generation looking at the redrawn map of Bosnia after
the Balkan war of 1912, Andric likewise wrote that they “saw nothing in those curving
lines, but they knew and understood everything, for their geography was in their blood
and they felt biologically their picture of the world” (p. 229). A balanced synopsis of
this history that places it in the context of the 20th-century developments that led up
to the 1999 Kosovo crisis is presented in William W. Hagen, “The Balkans’ Lethal Na-
tionalisms,” Foreign Affairs, July/August 1999, pp. 52–64. For more on this back -
ground, see Ivo H. Daalder and Michael E. O’Hanlon, Winning Ugly: NATO’s War to
Save Kosovo, Washington, D.C., Brookings Institution, 2000, pp. 1–100. See also
Michael Ignatieff, Virtual War: Kosovo and Beyond, New York, Metropolitan Books,
2000, especially pp. 11–65; Misha Glenny, The Balkans: Nationalism, War and the
Great Powers, 1809–1999, New York, Penguin Books, 2000; and Tim Judah, Kosovo:
War and Revenge, New Haven, Connecticut, Yale University Press, 2000.
5
6 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Federation began during the waning years of the cold war. 2 Under
the iron rule of Yugoslavia’s independent communist leader, Marshal
Josip Broz Tito, Kosovo had remained an autonomous and self-
governing province of Serbia for nearly 40 years, and members of its
largely ethnic Albanian populace were able to live a reasonable
approximation of normal lives. Once communist rule began to un-
ravel in the late 1980s after Tito’s death, however, the Serb minority
in Kosovo reacted forcefully against what they perceived to be willful
discrimination against them by the Kosovar Albanian authorities.
______________
2Today, Serbia and Montenegro (the latter is semiautonomous) are all that remain of
the former Yugoslavia.
Prelude to Combat 7
The trigger event that finally spurred the Clinton administration into
action with respect to Kosovo occurred on January 15, 1999, when
8 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
MUP and Serb paramilitary troops in hot pursuit of KLA fighters en-
tered the village of Racak and proceeded to slaughter 45 hapless
ethnic Albanian civilians. Ambassador Walker personally traveled to
Racak the next day to view the carnage, calling it a crime against hu-
manity and all but blaming Milosevic by name for having ordered it.
Two days later, the Yugoslav government, in response, declared
Walker persona non grata and issued an expulsion order, which
Walker ignored.3 The Racak massacre signaled the beginning of the
end to any further active role for the OSCE monitors, who were now
increasingly at physical risk themselves and who were ultimately
withdrawn less than a week before the commencement of Operation
Allied Force. It turned out to be a serious miscalculation on Bel-
grade’s part. What Milosevic may have thought was “just another
village” proved to be one too many as far as the United States and
NATO were concerned.4 The event marked the beginning of the final
countdown toward NATO’s ultimate decision to proceed with Allied
Force. On January 30, the NAC approved the launching of NATO air
attacks against Serbia if the Serb leaders continued to refuse nego-
tiations with their Kosovar counterparts.
______________
3The expulsion order was later rescinded by Milosevic.
4On this point, NATO’s Secretary General, Javier Solana, remarked that a Serb diplo-
mat had been heard to cite a rule of thumb to the effect that “a village a day would
keep NATO away.”
Prelude to Combat 9
At that point, with the gauntlet thrown down by Holbrooke, U.S. of-
ficials presented NATO’s ambassadors with a final proposed bomb-
ing plan against Serbia, the declared goals of which were a verifiable
halt to ethnic cleansing and atrocities on the ground in Kosovo; a
withdrawal of all but a token number of VJ, MUP, and paramilitary
troops from Kosovo; the deployment of an international peacekeep-
ing force in Kosovo; the return of refugees and their unhindered ac-
cess to aid; and the laying of groundwork for a future settlement in
Kosovo along the lines of the Rambouillet terms of reference.6
Commenting on the threatened campaign, the Supreme Allied
Commander in Europe (SACEUR), U.S. Army General Wesley Clark,
warned that “if required, we will strike in a swift and severe fashion.”
General Klaus Naumann, the chairman of NATO’s Military Commit-
tee, added that Milosevic was “severely mistaken” if he believed that
NATO would engage merely in pinprick attacks and then await his
response. 7
______________
5 In his exchange with Milosevic, Holbrooke said: “You understand our position?”
Milosevic: “Yes.” Holbrooke: “Is it absolutely clear what will happen when we leave,
given your position?” Milosevic: “Yes, you will bomb us. You are a big and powerful
nation. You can bomb us if you wish.” Bruce W. Nelan, “Into the Fire,” Time, April 5,
1999, p. 35. Later, Holbrooke added that Milosevic was “tricky, evasive, smart, and
dangerous,” further noting that his mood in the final confrontation was “calm, almost
fatalistic, unyielding.” “‘He Was Calm, Unyielding,’” Newsweek, April 5, 1999, p. 37.
6Jane Perlez, “Holbrooke to Meet Milosevic in Final Peace Effort,” New York Times,
March 22, 1999.
7R. Jeffrey Smith, “Belgrade Rebuffs Final U.S. Warning,” Washington Post, March 23,
1999.
Prelude to Combat 11
Two closely related U.S. joint task force (JTF) planning efforts called
Operations Flexible Anvil (commanded by U.S. Navy Vice Admiral
Daniel Murphy, commander of the 6th Fleet) and Sky Anvil
(commanded by USAF Lieutenant General Michael Short, comman-
der of the 16th Air Force at Aviano Air Base, Italy) followed in the
summer of 1998.10 Those efforts were terminated when Milosevic
initially agreed to a cease-fire after his October 5–13 talks with
______________
8Charles Babington and Helen Dewar, “President Pleads for Support,” Washington
Post, March 24, 1999.
9Telephone conversation with Lieutenant General Michael Short, USAF (Ret.), August
22, 2001.
10Flexible Anvil was a U.S.-only option that envisaged only ship-launched Tomahawk
and conventional air-launched cruise missile attacks over a 48- to 72-hour period,
roughly along the lines of Operation Desert Fox conducted against Iraq the following
December. Sky Anvil envisaged follow-on air strikes in a transition to a NATO opera-
tion (or an operation involving a more truncated coalition of the willing). General
Short believed that it was counterproductive to fragment these closely connected op-
tions into two separate plans, but he and Admiral Murphy were well acquainted and
kept each other informed. Conversation with Vice Admiral Daniel J. Murphy, USN,
commander, 6th Fleet, aboard the USS LaSalle, Gaeta, Italy, June 8, 2000.
12 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
11Lieutenant Colonel L. T. Wight, USAF, “What a Tangled Web We Wove: An After-
Action Assessment of Operation Allied Force’s Command and Control Structure and
Processes,” unpublished paper, p. 1.
12General John Jumper, USAF, testimony to the Military Readiness Subcommittee,
House Armed Services Committee, Washington, D.C., October 26, 1999. The most
fully developed of these iterations, called Operation Allied Talon, was a true phased air
campaign plan rooted in effects-based targeting and aimed at achieving concrete mili-
tary objectives. Despite the best efforts of the JTF Noble Anvil leadership (Admiral
James Ellis, General Jumper, and General Short) to sell this plan to SACEUR, General
Clark never adopted it. Instead, he elected to cut and paste different elements of the
different plans that he thought were most appropriate and labeled the resultant
product Operation Allied Force. Comments on an earlier draft by Hq USAFE/SA, April
6, 2001. General Jumper himself later confirmed that Allied Talon was a nonstarter.
Conversation with General John P. Jumper, USAF, Hq Air Combat Command, Langley
AFB, Virginia, May 15, 2001.
Prelude to Combat 13
NATO’s final plan was conceived from the start as a coercive opera-
tion only, with the implied goal of inflicting merely enough pain to
persuade Milosevic to capitulate. Its first phase, against only 51 ap-
proved integrated air defense system (IADS) targets and 40 approved
punishment targets out of 169 in NATO’s Master Target File, entailed
attacks against a combination of enemy air defenses and fixed army
installations that aimed at softening up Yugoslavia’s IADS and
demonstrating NATO’s ability to conduct precise air attacks with a
minimum of unintended damage. The second phase envisaged at-
tacks against military targets mainly, though not exclusively, below
the 44th parallel, which bisected Yugoslavia well south of Belgrade
(see Figure 2.1). Only in the third phase, if need be, would the
RAND MR1365-2.1
AUSTRIA
HUNGARY
ROMANIA
Taszar
Aviano SLOVENIA
Ljubljana Novi Sad
Vicenza Zagreb CROATIA Danube
River
Belgrade
BOSNIA AND
HERZEGOVINA arallel
Cervia 44th p
BULGARIA
Sarajevo SERBIA
Nis
Sofia
KOSOVO
Amendola MONTENEGRO
Pristina
Podgorica
ADRIATIC
SEA Kukes Skopje
Rome MACEDONIA
ITALY Durres
Tirana
ALBANIA
Naples
Gioia del Colle Brindisi
GREECE
For his part, General Clark had called for punitive air strikes against
Yugoslavia as early as January 1999, in response to the Serb massacre
of 45 Kosovar Albanians near the town of Racak just days before.
Persistent pressures from within NATO to explore a diplomatic solu-
tion, however, outweighed that recommendation for the early use of
force. The resulting delay gave Milosevic time to bolster his forces,
disperse important military assets, hunker down for an eventual
bombing campaign, and lay the final groundwork for the ethnic
cleansing of Kosovo. Owing to that delay, NATO lost any element of
surprise that may otherwise have been available.14
In the end, Operation Allied Force came just 10 days short of NATO’s
50th anniversary. The Clinton administration did not seek a UN
Security Council resolution approving the air attack plan, since it
knew that Russia and China had both vowed to veto any proposal
calling for air strikes.15 NATO’s going-in expectation was that the
bombing would be over very quickly. Indeed, so confident were its
______________
13Charles Babington and William Drozdiak, “Belgrade Faces the 11th Hour, Again,”
Washington Post, March 22, 1999. For more first-hand comment on the intra-NATO
politics that preceded Allied Force, see General Wesley K. Clark, Waging Modern War:
Bosnia, Kosovo, and the Future of Combat, New York, Public Affairs, 2001, especially
pp. 121–189.
14William Drozdiak, “Politics Hampered Warfare, Clark Says,” Washington Post, July
20, 1999.
15This should not be taken to suggest that NATO’s air war against Serbia was a uni-
lateral action undertaken without regard for the UN whatsoever. On the contrary, in
March 1998 the Security Council had expressly recognized in Resolution 1160 that the
Serb government’s repression of the ethnic Albanian population in Kosovo constituted
a threat to international peace and security, a view later repeated in Resolution 1199
six months before the start of Allied Force, which called for action aimed at heading off
“the impending humanitarian catastrophe” in Kosovo. As an IISS comment later
noted, NATO’s air war for Kosovo thus constituted “a highly significant precedent,” in
that it established “more firmly in international law the right to intervene on
humanitarian grounds, even without an express mandate from the Security Council.”
Strategic Survey 1999/2000, London, England, The International Institute for Strategic
Studies, 2000, p. 26.
Prelude to Combat 15
______________
16Jane Perlez, “U.S. Option: Air Attacks May Prove Unpalatable,” New York Times,
March 23, 1999.
17Steven Erlanger, “U.S. Issues Appeal to Serbs to Halt Attack in Kosovo,” New York
Times, March 23, 1999.
This page intentionally blank
Chapter Three
THE AIR WAR UNFOLDS
______________
1“AWOS [Air War Over Serbia] Fact Sheet,” Hq USAFE/SA, December 17, 1999. See
also The Military Balance, 1998/99, London, International Institute for Strategic Stud-
ies, 1998, p. 100.
17
18 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
ably with selected upgrades, its operators knew U.S. tactics well and
had practiced air defense drills and honed their operational tech-
niques for more than four decades. They also had the benefit of
more equipment and better training than did the Bosnian Serbs in
1995. Finally, they enjoyed the advantage of being protected both by
mountainous terrain and by the cover of inclement weather when
the air war began.
______________
2Discussions with former East European strategic and tactical SAM operators on IADS
visual observer employment doctrine, as reported to the author by Hq USAFE/IN, May
18, 2001.
3John Diamond, “Yugoslavia, Iraq Talked Air Defense Strategy,” Philadelphia Inquirer,
March 30, 1999.
4Michael R. Gordon, “NATO to Hit Serbs from 2 More Sides,” New York Times, May 11,
1999. This last system featured the Bofors 40mm gun tied to the Giraffe radar-based
low-altitude air defense system (LAADS). It was the only radar-cued (as opposed to
radar-directed) AAA weapon fielded in the war zone and possibly the most potent low-
altitude AAA threat because of its local Giraffe-based LAADS command and control
system. Peter Rackham, ed., Jane’s C4I Systems, 1994–95, London, Jane’s Information
Group, 1994, p. 107.
The Air War Unfolds 19
______________
5 Paul Richter, “U.S. Pilots Face Perilous Task, Pentagon Says,” Los Angeles Times,
March 20, 1999. In testimony to the Senate Armed Services Committee on the eve of
the war, Ryan added: “I ran the air campaign in Bosnia, and this defensive array is
much more substantive . . . two or three times more so. It is deep and redundant.
Those guys [in Bosnia] were good, but these guys are better. There is a very real possi-
bility we will lose aircraft trying to take it on.” David Atkinson, “Stealth Could Play Key
Role in Kosovo, Despite Bad Weather,” Defense Daily, March 23, 1999, p. 1.
6Bruce W. Nelan, “Into the Fire,” Time, April 5, 1999, p. 31.
7Francis X. Clines, “NATO Opens Broad Barrage Against Serbs as Clinton Denounces
‘Brutal Repression,’” New York Times, March 25, 1999.
20 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Milosevic did not desist from his offensive in Kosovo.8 As noted ear-
lier, it was accepted as a given by the Clinton administration that
Milosevic would settle quickly. As Secretary of State Albright clearly
attested to this expectation in a television interview on the evening
that the air attacks began: “I don’t see this as a long-term opera-
tion.” 9
The air war commenced with 250 committed U.S. aircraft, including
120 land-based fighters, 7 B-52s, 6 B-2s, 10 reconnaissance aircraft,
10 combat search and rescue (CSAR) aircraft, 3 airborne command
and control center (ABCCC) aircraft, and around 40 tankers.10 As for
NATO’s additional 18 members, 13 contributed aircraft for use in the
operation, with 11 allies (Britain, France, Germany, Italy, Belgium,
Canada, Denmark, the Netherlands, Norway, Spain, and Turkey)
eventually participating in offensive and defensive air combat
operations of all types. The first wave of attacks on the night of
March 24 consisted of cruise missile launches only, featuring TLAMs
fired by four U.S. surface ships (including USS Gonzales and USS
Philippine Sea), two U.S. fast-attack submarines (USS Albuquerque
and USS Miami), and a British attack submarine (HMS Splendid)
operating in the Adriatic Sea. This initial wave further included
AGM-86C CALCMs launched against hardened enemy structures by
six B-52s flying outside Yugoslav airspace. The latter were the first
shots fired in the operation.11 The initial target hits occurred shortly
______________
8Paul Richter, “Time Is Not on the Side of U.S., Allies,” Los Angeles Times, March 25,
1999.
9John T. Correll, “Assumptions Fall in Kosovo,” Air Force Magazine, June 1999, p. 4.
10This study has taken special care to characterize Operation Allied Force as an “air
war” or an “air effort,” rather than as a full-fledged “air campaign.” Although that ef-
fort continues to be widely portrayed as the latter, formal Air Force doctrine defines an
air campaign as “a connected series of operations conducted by air forces to achieve
joint force objectives within a given time and area.” Air Force Basic Doctrine, Maxwell
AFB, Alabama, Hq Air Force Doctrine Center, AFDD-1, September 1997, p. 78. By that
standard, NATO’s air war for Kosovo did not attain to the level of a campaign, as did
the earlier Operations Desert Storm and Deliberate Force. Rather, it was a continu-
ously evolving coercive operation featuring piecemeal attacks against unsystemati-
cally approved targets, not an integrated effort aimed from the outset at achieving
predetermined and identifiable operational effects.
11The effectiveness of these initial standoff attacks was not impressive. During the
first two weeks, no B-52 succeeded in launching all eight of its CALCMs. In one in-
stance, six out of eight were said to have failed. Also, the two times that B-52s later
fired the AGM-142 Have Nap cruise missile, both launches were reportedly opera-
The Air War Unfolds 21
_____________________________________________________________
tional failures. See John D. Morrocco, David Fulghum, and Robert Wall, “Weather,
Weapons Dearth Slow NATO Strikes,” Aviation Week and Space Technology, April 5,
1999, p. 26, and William M. Arkin, “Kosovo Report Short on Weapons Performance
Details,” Defense Daily, February 10, 2000, p. 2.
12 An important qualification is warranted here. Although the opening-night ap-
proved aim points largely entailed fixed IADS targets, the limited attacks conducted
against them were not part of a phased campaign plan in which rolling back the en-
emy IADS was a priority. There was no strategic emphasis on IADS takedown in these
attacks. Comments on an earlier draft by Hq USAFE/IN, May 18, 2001.
13Italian bases used included Aviano, Gioio del Colle, Villafranca, Amendola, Cervia,
Gazzanise, Ghedi, Piacenza, Istrana, Falconara, Practica di Mare, Brindisi, and
Sigonella. German bases used were Royal Air Force (RAF) Bruggen, Rhein Main Air
Base (AB), Spangdahlem AB, and Ramstein AB. Bases made available by the United
Kingdom were RAF Fairford, RAF Lakenheath, and RAF Mildenhall. Spain provided
Moron AB, and France provided Istres. For a complete list of all participating allied air
assets, their units, and their bases, as well as a tabulation of the Yugoslav IADS and air
order of battle as of April 20, see Benoit Colin and Rene J. Francillon, “L’OTAN en
Guerre!” Air Fan, May 1999, pp. 12–19. See also John E. Peters, Stuart Johnson, Nora
Bensahel, Timothy Liston, and Traci Williams, European Contributions to Operation
Allied Force: Implications for Transatlantic Cooperation, Santa Monica, California,
RAND, MR-1391-AF, 2001.
14Robert Hewson, “Operation Allied Force: The First 30 Days,” World Air Power Jour-
nal, Fall 1999, p. 16.
22 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
tude floor (or “hard deck”) of 15,000 ft to remain above their killing
envelopes.
Targets attacked the first night were reviewed with special care at the
White House by President Clinton, Secretary of Defense William Co-
hen, and General Shelton. Some proposed targets were removed
from the list by dissenting NATO leaders out of concern for causing
collateral damage because of their close proximity to civilian build-
ings. In other borderline cases in which targets were reluctantly ap-
proved, the recommended bomb size was reduced to minimize or
preclude collateral damage. One of every five laser-guided bombs
dropped by an F-117 the first night was a 500-lb GBU-12 instead of a
2,000-lb GBU-27. That meant less likelihood of the bomb’s causing
inadvertent collateral damage, but also a lower probability of de-
stroying the intended target. The rules of engagement were un-
compromisingly restrictive, with pilots instructed to return home
with their weapons unless their assigned target could be positively
identified.15
In all, some 400 sorties were flown the first night, including 120 strike
missions against 40 targets consisting of five airfields, five army gar-
risons, communications centers, and storage depots, in addition to
IADS facilities. Only a few SA-3 and SA-6 SAMs were launched
against attacking NATO aircraft the first night. All the same, Pen-
tagon officials anticipated the day after that at least a dozen NATO
aircraft losses could be incurred should the operation continue be-
yond just a few days.16 Contrary to early Western press reports, Serb
IADS operators never intentionally husbanded their SAMs. Instead,
after experiencing allied SEAD operations for the first time, they
adapted their tactics to balance lethality with survivability, with the
result that they were always present and aggressive—even as they
showed greater firing discipline than the Iraqis did during Desert
Storm.17
______________
15Nelan, “Into the Fire,” p. 32.
16Steven Lee Myers, “Early Attacks Focus on Web of Air Defense,” New York Times,
March 25, 1999.
17Comments on an earlier draft by Hq USAFE/IN, May 18, 2001.
The Air War Unfolds 23
Attacks carried out by NATO aircrews the second night were de-
scribed as “significantly heavier” than those the first night. Targets
included the VJ barracks at Urosevac and Prizren in Kosovo; the mili-
tary airfields at Nis in southern Serbia and Golubovci near Podgorica,
Montenegro; and other Serb military facilities near Trstenik and
Danilovgrad. 20 That night, fewer than 10 SAMs were fired, none of
which succeeded in scoring a hit. The third afternoon, a USAF F-15C
downed two more MiG-29s, which evidently had lost contact with
their ground controller and inadvertently strayed into Bosnian
airspace. Although their intended NATO targets were never posi-
tively determined, it was the subsequent conclusion of the allied air
commander, Lieutenant General Short, that the Serb pilots had sim-
ply lost any semblance of air situation awareness and, as a result, set
themselves up as easy prey for the F-15. 21
______________
18This suggested that the Serb IADS may have been unable to deconflict its SAMs and
fighters operating in the same airspace because of identification and discrimination
problems.
19 Barton Gellman, “Key Sites Pounded for 2nd Day,” Washington Post, March 26,
1999. See also John D. Morrocco and Robert Wall, “NATO Vows Air Strikes Will Go the
Distance,” Aviation Week and Space Technology, March 29, 1999, p. 34.
20Hewson, “Operation Allied Force,” p. 17.
21 Telephone conversation with Lieutenant General Michael Short, USAF (Ret.),
August 16, 2001.
24 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
22Hewson, “Operation Allied Force,” p. 18.
23John F. Harris, “Clinton Saw No Alternative to Airstrikes,” Washington Post, April 1,
1999.
24Johanna McGeary, “The Road to Hell,” Time, April 12, 1999, p. 42.
25For an informed treatment of the KLA and its origins, goals, and prospects by The
New York Times’ Balkans bureau chief from 1995 to 1998, see Chris Hedges, “Kosovo’s
Next Masters?” Foreign Affairs, May/June 1999, pp. 24–42.
The Air War Unfolds 25
______________
26 Carla Anne Robbins, Thomas E. Ricks, and Neil King, Jr., “Milosevic’s Resolve
Spawned Unity, Wider Bombing List in NATO Alliance,” Wall Street Journal, April 27,
1999. Unlike nearly all other NATO principals, General Naumann cautioned even be-
fore the air war began that although the intent was to be quick, Operation Allied Force
could well turn out to be “long and protracted.” Paul Richter and John-Thor
Dahlburg, “NATO Broadens Its Battle Strategy,” Los Angeles Times, March 24, 1999.
27William Drozdiak, “NATO Leaders Struggle to Find a Winning Strategy,” Washington
Post, April 1, 1999.
26 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
the Greek and Italian representatives, who had been calling for an
Easter bombing pause in the hope that it might lead to negotia-
tions.28
NATO went into this second phase earlier than anticipated because
of escalating Serb atrocities on the ground. Up to that point, the air
attacks had had no sought-after effect on Serb behavior whatsoever.
On the contrary, Serbia’s offensive against the Kosovar Albanians
intensified, with Serb troops burning villages, arresting dissidents,
and executing supposed KLA supporters. The Serbs continued un-
opposed in their countercampaign of ethnic cleansing, ultimately
forcing most of the 1.8 million ethnic Albanians in Kosovo from their
homes.29
______________
28The latter rumblings prompted concern in U.S. and NATO military circles that once
any such pause might be agreed to, it would be that much more difficult to resume the
bombing after the pause had expired. In the end, no pause in the bombing occurred
at any time during Allied Force, other than those occasioned by bad weather.
29 After careful examination, the provision of airlift relief missions for the Kosovar
refugees was ruled out by U.S. and NATO planners because they were deemed exces-
sively dangerous in the face of threats from enemy ground fire and because of concern
that any delivered supplies would end up in the wrong hands.
30Bradley Graham and William Drozdiak, “Allied Action Fails to Stop Serb Brutality,”
Washington Post, March 31, 1999.
31Craig R. Whitney, “On 7th Day, Serb Resilience Gives NATO Leaders Pause,” New
York Times, March 31, 1999.
The Air War Unfolds 27
Five B-1Bs were added to the U.S. Air Force’s bomber contingent at
the start of the second week. In preparing them for combat, what
normally would have taken months of effort to program the aircraft’s
mission computers was compressed into fewer than 100 hours dur-
ing a single week as Air Force officers and contractors updated the
computers with the latest intelligence on enemy radar and SAM
threats. One aircraft with the latest updated software installed, the
Block D upgrade of the Defensive System Upgrade Program, passed a
critical flight test at the 53rd Wing at Eglin AFB, Florida, and two B-1s
were committed to action over Yugoslavia two days later. These air-
craft, which, alongside the B-52s, operated out of RAF Fairford in
England, employed the Raytheon ALE-50 towed decoy to good effect
for the first time in combat.32 They were still test-configured aircraft
flown by test crews. The B-1s, all test-configured with Block D up-
grades, typically flew two-ship missions against military area targets,
such as barracks and marshaling yards.
______________
32Of 10 known Serb SAMs that reportedly guided on B-1s during the course of the air
war, all were believed to have been successfully diverted to the decoys. See David
Hughes, “A Pilot’s Best Friend,” Aviation Week and Space Technology, May 31, 1999, p.
25. The commander of USAFE, General John Jumper, later explained the Serb IADS
tactic employed: Radars in Montenegro would acquire and track the B-1s as they flew
in from over the Adriatic Sea, arced around Macedonia, and proceeded north into
Kosovo. Those acquisition radars would then hand off their targets to SA-6s, whose
radars came up in full target-track mode and fired the missiles, which headed straight
for the ALE-50 and took it out, just as the system was designed to work. “Jumper on
Air Power,” Air Force Magazine, July 2000, p. 43.
28 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
strike sorties that had been conducted during the first 12 hours of
Desert Storm.
______________
33Bradley Graham, “Bombing Spreads,” Washington Post, March 29, 1999.
34The USAF flew 84 percent of those sorties, the NATO allies 10 percent, and the U.S.
Navy 6 percent.
35Graham, “Bombing Spreads.”
36Craig R. Whitney, “NATO Had Signs Its Strategy Would Fail Kosovars,” New York
Times, April 1, 1999.
37William Drozdiak and Bradley Graham, “NATO Frustration Grows as Mission Falls
Short,” Washington Post, April 8, 1999.
The Air War Unfolds 29
______________
38 Thomas W. Lippman and Dana Priest, “NATO Builds Firepower for 24-Hour At-
tacks,” Washington Post, March 30, 1999.
39The NAC did not formally approve strikes on Phase III targets per se, although it did
assent to target classes within Phase III.
40Hewson, “Operation Allied Force,” p. 22.
30 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
U.S. naval aviation, unavailable for the initial phase of Operation Al-
lied Force, joined the fray when the aircraft carrier USS Theodore
Roosevelt arrived on station in the Ionian Sea south of Italy two weeks
afterward, on April 6. The air wing assigned to the Theodore Roo-
sevelt flew complete and self-sustaining strike packages, including
F-14Ds and F/A-18s for surface-attack operations, EA-6Bs for the
suppression of enemy air defenses, F-14s in the role of airborne
forward air controllers, and E-2Cs performing as ABCCC platforms.
These packages typically flew missions only against dispersed and
hidden enemy forces in Kosovo, although on one occasion, on April
15, they struck a hardened aircraft bunker at the Serbian air base at
Podgorica in Montenegro in the first of several allied efforts to
neutralize a suspected air threat against the U.S. Army’s Task Force
Hawk deployed in Albania (see below).41 The E-2C, normally
operated as an airborne early warning (AEW) platform to screen the
carrier battle group from enemy air threats, was used in Allied Force
to provide an interface between the CAOC and naval air assets
______________
41 Conversation with Vice Admiral Daniel J. Murphy, USN, commander, 6th Fleet,
aboard USS LaSalle, Gaeta, Italy, June 8, 2000. See also Vice Admiral Daniel J. Murphy,
USN, “The Navy in the Balkans,” Air Force Magazine, December 1999, p. 49. According
to a later account by General Jumper, the strike against the Podgorica airfield was the
most concentrated effort placed on any target throughout the entire course of Allied
Force. To satisfy SACEUR’s objective, General Short needed to neutralize the airfield’s
sortie generation capacity completely. At the time the target was selected, only 50 per-
cent of the aim points required to meet that objective had been identified. It took 48
hours to accomplish the additional target analysis and to free up additional required
NATO assets to carry out this strike. Since the Theodore Roosevelt had just arrived in
the theater, it had not been tasked in the April 15 Air Tasking Order and accordingly
had assets that were immediately available. As a result, F-14 and F/A-18 aircraft struck
the hardened aircraft bunker (the highest-value critical element) and used CAOC
(Combined Air Operations Center) assets to assist in targeting and weaponeering.
Other NATO assets struck the remaining critical elements 48 hours later and met
SACEUR’s objectives. Conversation with General John P. Jumper, USAF, Hq Air
Combat Command, Langley AFB, Virginia, May 15, 2001.
The Air War Unfolds 31
It was hard during the first few weeks for outside observers to assess
and validate the Pentagon’s and NATO’s claims of making progress
because U.S. and NATO officials had so deliberately refrained from
disclosing any significant details about the operation. Instead, ad-
ministration and NATO sources limited themselves to vague general-
izations about the air war’s effects, using such hedged terms as
“degrading,” “disrupting,” and “debilitating” rather than the more
unambiguous “destroying.” On this studiously close-mouthed pol-
icy, the Defense Department’s spokesman, Kenneth Bacon, declared
that a precedent was being intentionally set, since both Secretary
Cohen and General Shelton had seen a need to “change the culture
of the Pentagon and make people more alert to the dangers that can
flow from being too generous—or you could say profligate or lax—
with operational details.”43
______________
42See Commander Wayne D. Sharer, USN, “The Navy’s War over Kosovo,” Proceed-
ings, U.S. Naval Institute, October 1999, pp. 26–29; and Robert Wall, “E-2Cs Become
Battle Managers with Reduced AEW Role,” Aviation Week and Space Technology, May
10, 1999, p. 38.
43Jason DeParle, “Allies’ Progress Remains Unclear as Few Details Are Made Public,”
New York Times, April 5, 1999.
44In fairness to that effort, however, and given the many constraints that affected it—
in contrast to the far fewer constraints that affected Desert Storm—weather, mainly an
irritant during the Gulf War, was a significant factor during Operation Allied Force.
Bad weather, combined with the higher population density of Serbia, the concern for
collateral damage, and the increased surface-to-air threat, could easily have con-
tributed to a lower relative intensity of strike operations. I thank Major Richard
Leatherman, Hq Air Force Doctrine Center, for having called this possibility to my at-
tention.
32 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
By the third week, NATO’s strategic goals had shifted from seeking to
erode Milosevic’s ability to force an exodus of Kosovar Albanian
civilians to enforcing a withdrawal of Serb forces from Kosovo and a
return of the refugees home. That shift in strategy was forced by
Milosevic’s early seizure of the initiative and his achievement of a
near–fait accompli on the ethnic cleansing front. Up to that point,
President Clinton had merely insisted that the operation’s goal was
to ensure that Milosevic’s military capability would be “seriously
diminished.”45
______________
45John M. Broder, “Clinton Says Milosevic Hurts Claim to Kosovo,” New York Times,
March 31, 1999.
46“Hope for the Best, and a Spot of Golf,” The Economist, April 3, 1999, p. 9.
The Air War Unfolds 33
By the end of the third week, in large measure out of frustration over
the operation’s continued inability to get at the dug-in and elusive VJ
positions in Kosovo, Clark requested a deployment of 300 more air-
craft to support the effort. That request, which would increase the
total number of committed U.S. and allied aircraft to nearly 1,000,
entailed more than twice the number of allied aircraft (430) on hand
when the operation began on March 24—and almost half of what the
allied coalition had had available for Desert Storm. For the United
States, it represented a 60 percent increase over the 500 U.S. aircraft
already deployed (see Figure 3.1 for the ultimate proportions of U.S.
and allied aircraft provided to support Allied Force). Among other
things, it prompted understandable concern about where to base
RAND MR1365-3.1
USN/USMC
5%
The call for 300 additional aircraft followed on the heels of an earlier
request by Clark for 82 more aircraft, which had been promptly ap-
proved by the Pentagon. This time, Pentagon officials expressed
surprise at the size of Clark’s request and openly questioned whether
it would be approved in its entirety. 48 The principal concern was
that it would draw precious assets, notably such low-density/high-
demand aircraft as the E-3 airborne warning and control system
(AWACS) and EA-6B Prowler, away from Iraq and Korea. The service
chiefs reportedly complained that Clark’s requested quantities repre-
sented a clear case of overkill and that USEUCOM was not making
the most of the forces already at its disposal.
In addition, Clark asked for the USS Enterprise and its 70-aircraft air
wing, which would necessitate extending the carrier’s cruise length
and thereby breaking a firm Navy rule of not keeping aircrews and
sailors at sea for any longer than six months at a single stretch. The
request was opposed by the chief of naval operations, Admiral Jay
Johnson, and Secretary of the Navy Richard Danzig. In the end, the
Enterprise was made available by the Navy for diversion to the
Adriatic as requested, but its air wing was never tasked by the CAOC,
and it never participated in Allied Force. Once the additional aircraft
were approved, NATO asked Hungary to make bases available and
Turkey to help absorb those aircraft. Figure 3.2 shows how the in-
theater buildup of aircraft ultimately played itself out.
As a part of his requested force increment, Clark also asked for a de-
ployment of Army AH-64 Apache attack helicopters. Although the
other aircraft were eventually approved by Secretary Cohen and the
Joint Chiefs of Staff (JCS), this particular request was initially
______________
47 Steven Lee Myers, “Pentagon Said to Be Adding 300 Planes to Fight Serbs,” New
York Times, April 13, 1999.
48One former senior U.S. officer commented that Clark had presented “a wish list that
would choke a horse.” Elaine M. Grossman, “Clark’s Firepower Request for Kosovo
Prompts Anxiety Among Chiefs,” Inside the Pentagon, April 15, 1999.
The Air War Unfolds 35
RAND MR1365-3.2
1,200
1,000
800
Number of aircraft
USAF
600
400
USA, USN,
USMC
200
Allies
0
February 23
January 1 Rambouillet March 24 April 23 June 10
talks end Allied NATO summit Cease-fire
Force
begins
SOURCE: Hq USAFE/SA.
In the end, despite Army and JCS reluctance, Clark prevailed in his
request for the Apaches and announced that 24 would be deployed
to Albania from their home base at Illesheim, Germany. Pentagon
spokesmen went out of their way to stress that the Apaches were in-
tended solely as an extension of the air effort and not as an implied
36 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
49Bradley Graham and Dana Priest, “Allies to Begin Flying Refugees Abroad,” Wash-
ington Post, April 5, 1999. In his subsequent memoirs, Clark frankly excoriated what
he called “the reluctant Army mind-set in Washington” on this issue and dismissed
other critics of his requested AH-64 commitment as “the voices of conventional air
power” coming from “commanders who had no experience with the Apaches.” Gen-
eral Wesley K. Clark, Waging Modern War: Bosnia, Kosovo, and the Future of Combat,
New York, Public Affairs, 2001, pp. 230, 278.
50To which General Naumann countered in resigned exasperation: “It isn’t a good
Rembrandt.” Robbins, Ricks, and King, Jr., “Milosevic’s Resolve Spawned More Unity
in Alliance and a Wider Target List.”
51Michael Ignatieff, “The Virtual Commander: How NATO Invented a New Kind of
War,” The New Yorker, August 2, 1999, p. 32.
The Air War Unfolds 37
As the air war entered its fifth week, Clark admitted that Milosevic
was still pouring reinforcements into Kosovo continuously and that
“if you actually added up what’s there on a given day, you might find
that he’s strengthened his forces in there.” 53 Much as during some
periods of Desert Storm, adverse weather at the five-week point had
forced a cancellation or failure of more than half of all scheduled
bombing sorties on 20 of the first 35 days of air attacks. Seemingly
resigned to a waiting game as the air war appeared stalled after more
than a month of continual bombing, a senior NATO diplomat con-
fessed that it now felt as though Operation Allied Force “had been
put on autopilot. Now we are basically waiting for something to
crack in Belgrade.”54 In light of the stalled offensive, some saw the
air war now threatening to stretch into summer 1999, if not longer.
______________
52Bradley Graham and John Lancaster, “Most NATO Bombing Raids Target Previously
Hit Sites,” Washington Post, April 21, 1999. In fairness to NATO planners, some of
those reattacks were valid, because a few especially large area targets entailed nu-
merous individual aim points, some of which were missed in the initial attacks. The
vast majority, however, merely entailed what many frustrated NATO crewmembers re-
ferred to as “bouncing rubble,” having no practical effect and presenting considerable
added risk to their own survivability. It was not uncommon for aircrews to complain
vocally about having their “warm bodies sent out all over again to turn bricks into
powder.”
53Craig R. Whitney, “NATO Chief Admits Bombs Fail to Stem Serb Operations,” New
York Times, April 28, 1999.
54Neil King, Jr., “War Against Yugoslavia Lapses into Routine, but Clock Is Ticking,”
Wall Street Journal, May 6, 1999.
38 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
55Doyle McManus, “Clinton’s Massive Ground Invasion That Almost Was,” Los Ange-
les Times, June 9, 2000.
56General John Jumper, USAF, “Oral Histories Accomplished in Conjunction with
Operation Allied Force/Noble Anvil.”
57Dana Priest, “Target Selection Was Long Process,” Washington Post, September 20,
1999. One must take care, however, not to confuse Master Target File growth with
approved target growth. Although target nominations increased dramatically as the
air war entered full swing, getting those targets individually approved remained a
challenge throughout the air war to the very end.
The Air War Unfolds 39
______________
58 Eric Schmitt and Steven Lee Myers, “NATO Said to Focus Raids on Serb Elite’s
Property,” New York Times, April 19, 1999.
59In the end, however, only some 10 percent of the 48th Fighter Wing’s F-15E combat
missions were flown out of Lakenheath. The remainder were flown out of the wing’s
40 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
_____________________________________________________________
forward operating location at Aviano. Conversation with USAF F-15E aircrews, 492nd
Fighter Squadron, RAF Lakenheath, England, April 28, 2001.
60Robert Hewson, “Allied Force, Part II: Overwhelming Air Power,” World Air Power
Journal, Winter 1999/2000, p. 99.
61Paul Richter, “Bunker-Busters Aim at Heart of Leadership,” Los Angeles Times, May
5, 1999.
62Ibid. The Pentagon’s formal report to Congress later indicated that “some” hard-
ened underground command bunkers had been destroyed.
63 Dana Priest, “France Acted as Group Skeptic,” Washington Post, September 20,
1999, and David A. Fulghum, “Russians Analyze U.S. Blackout Bomb,” Aviation Week
and Space Technology, February 14, 2000, p. 59.
The Air War Unfolds 41
Whatever the case, the attack moved NATO over a new threshold and
brought the war, for the first time, directly to the Serbian people. By
the end of the seventh week, there began to be reports of Yugoslav
officials openly admitting that the country was on the verge of
widespread hardship because of the air war’s mounting damage to
the nation’s economy, which had already been weakened by almost
four years of international sanctions imposed for Serbia’s earlier role
in the war in Bosnia.67 The destruction of one factory in Krujevac
that produced automobiles, trucks, and munitions resulted in 15,000
people being put out of work, plus 40,000 more who were employed
by the factory’s various subcontractors. Attacks against other facto-
ries had similar effects on the Yugoslav economy. By the time Allied
Force had reached its halfway point, the bombing of infrastructure
targets had halved Yugoslavia’s economic output and deprived more
______________
64An inertial navigation system (INS)–guided version of the weapon, a variant of the
wind-corrected munitions dispenser, is now said to be entering the U.S. munitions in-
ventory. Fulghum, “Russians Analyze U.S. Blackout Bomb.”
65David A. Fulghum, “Electronic Bombs Darken Belgrade,” Aviation Week and Space
Technology, May 10, 1999, p. 34.
66The results were more symbolic than strategically significant. After the May 3 at-
tack, some 500 workers managed to clear the filaments sufficiently to restart the
equipment within 15 hours. After a similar attack on May 8, the threads were cleared
within 4 hours. William Arkin, “Smart Bombs, Dumb Targeting?” Bulletin of the
Atomic Scientists, May/June 2000, p. 52.
67 Robert Block, “In Belgrade, Hardship Grows Under Sustained Air Assault,” Wall
Street Journal, May 12, 1999.
42 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Only during the last two weeks of Allied Force, however, did NATO
finally strike with real determination against Serbia’s electrical power
generating capability, a target set that had been attacked in Baghdad
from the very first days of Desert Storm. The earlier “soft” attacks at
the beginning of May with graphite filament bombs against the
transformer yards of Yugoslavia’s main power grid had caused a
temporary disruption of the power supply by shorting out trans-
formers and disabling them rather than destroying them. But this
time, in perhaps the single most attention-getting strike of the entire
air war up to that point, the Yugoslav electrical grid was severely
damaged over the course of three consecutive nights starting on May
24. Those attacks, directed against electrical power facilities and re-
lated targets in Belgrade, Novi Sad, and Nis, the three largest cities in
Serbia, shut off the power to 80 percent of Serbia, leaving millions
without electricity or water service. They affected the heart of Yu-
goslavia’s IADS, as well as the computers that ran its banking system
and other important national consumers of electricity.69
______________
68Steven Erlanger, “Economists Find Bombing Cuts Yugoslavia’s Production in Half,”
New York Times, April 30, 1999.
69Philip Bennett and Steve Coll, “NATO Warplanes Jolt Yugoslav Power Grid,” Wash-
ington Post, May 25, 1999.
The Air War Unfolds 43
Indeed, even before the operation was a week old, indications had
begun to mount that senior administration officials were starting to
have second thoughts about the advisability of having peremptorily
______________
70Ibid.
71William Drozdiak, “Allies Target Computer, Phone Links,” Washington Post, May 27,
1999.
72Ignatieff, “The Virtual Commander,” p. 35.
73James Gerstenzang and Elizabeth Shogren, “Serb TV Airs Footage of 3 Captured U.S.
Soldiers,” Los Angeles Times, April 1, 1999.
74On that account, Clark later acknowledged that his air commanders were no hap-
pier than he was with the absence of a ground threat, noting that it was “sort of an un-
natural act for airmen to fight a ground war without a ground component.” Ignatieff,
“The Virtual Commander,” p. 33.
44 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
ruled out a ground option before launching into Allied Force. The
chairman of the JCS, General Shelton, for example, remarked that
there were no NATO plans “right now” to introduce ground troops
short of a peace settlement in Kosovo.75 In a similarly hedged re-
mark, Secretary Cohen pointed out that the Clinton administration
and NATO had no plans to introduce any ground troops “into a hos-
tile environment,” leaving open the possibility that they might con-
template putting a ground presence into a Kosovo deemed
“nonhostile” before the achievement of a settlement. 76 By the end of
the second week, Secretary of State Albright went further yet toward
hinting at the administration’s growing discomfiture over having
ruled out a ground threat when she allowed that NATO might change
its position and put in ground troops should the bombing succeed in
creating a “permissive environment.”77 Ultimately, the air war’s
continued indecisiveness led President Clinton himself to concede
that he would consider introducing ground troops if he became
persuaded that the bombing would not produce the desired
outcome. In a clear contradiction to his earlier position on the issue,
he asserted that he had “always said that . . . we have not and will not
take any option off the table.” That statement was later described by
a U.S. official as testimony to an ongoing administration effort “to
break out of a rhetorical box that we should never have gotten
into.” 78
______________
75 Paul Richter, “Use of Ground Troops Not Fully Ruled Out,” Los Angeles Times,
March 29, 1999.
76Rowan Scarborough, “Military Experts See a Need for Ground Troops,” Washington
Times, March 30, 1999.
77Rowan Scarborough, “Momentum for Troops Growing,” Washington Times, April 5,
1999.
78John F. Harris, “Clinton Says He Might Send Ground Troops,” Washington Post, May
19, 1999. In an earlier attempt at revisionism, Secretary of State Albright upbraided an
interviewer by flatly declaring that “we never expected this to be over quickly,” in
complete contradiction to her categorical pronouncement the first night of the air war
11 days earlier that “I think that this is something, the deter and damage, is something
that is achievable within a relatively short time.” John Harris, “Reassuring Rhetoric,
Reality in Conflict,” Washington Post, April 8, 1999.
The Air War Unfolds 45
At the same time, a pronounced rift emerged between Clark and his
Pentagon superiors over Clark’s insistence on replacing talk with de-
termined action in connection with preparations for a ground inva-
sion. In his memoirs, Clark later gave candid vent to his frustration
over this rift when he referred to the “divide between those in
Washington who thought they understood war and those [of us] in
Europe who understood Milosevic, the mainsprings of his power,
and the way to fight on this continent.”80 Earlier in April, he had
challenged U.S. and British officers at NATO headquarters to con-
sider “what if” options for a potential ground war. Out of frustration
over the refusal of both Washington and his NATO masters to coun-
tenance any serious consideration of a ground component to Allied
Force, he also asked the Army, shortly after the air war began, to send
him a half-dozen officers from the School of Advanced Military
Studies at Fort Leavenworth, Kansas, to draw up secret plans for a
broad spectrum of ground options, ranging from sending in peace-
keepers to police any settlement that might be achieved single-
handedly by the air war to launching a full-fledged, opposed-entry
land invasion if all else proved wanting. It soon became clear from
______________
79 Dan Balz, “U.S. Consensus Grows to Send in Ground Troops,” Washington Post,
April 6, 1999.
80Clark, Waging Modern War, p. 303. As a testament to the depth of his conviction on
the criticality of getting serious about laying the groundwork for a land invasion, Clark
in mid-May wrote a letter to Secretary General Solana which, he said, “demonstrated
at length how moving into ground-force preparations would exponentially increase
[NATO’s] leverage against Milosevic.” Ibid, pp. 307–308, emphasis added.
46 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
that inquiry that only about a dozen roads led into Kosovo from Al-
bania. Like Kosovo’s bridges, they were heavily mined and strongly
defended, with VJ troops well positioned on the high ground of the
most strategically crucial terrain. Accordingly, the study concluded
that the best invasion routes would be from Hungary and Croatia
into the flatter terrain of northern Serbia.
______________
81Interview by RAND staff, Washington, D.C., June 11, 2000. The UK Ministry of De-
fense’s director of operations in Allied Force, Air Marshal Sir John Day, however, later
commented that there was never much military enthusiasm for a double envelopment
through Hungary. Conversation with Air Marshal Day, RAF Innsworth, United King-
dom, July 26, 2000.
82Thomas W. Lippman and Bradley Graham, “NATO Chief Asks Review of Invasion
Planning,” Washington Post, April 22, 1999.
The Air War Unfolds 47
By most accounts, the turning point in facing up to the need for a se-
rious ground option came on May 27, when Cohen met secretly in
Bonn with his four principal NATO counterparts, the British, French,
German, and Italian defense ministers, in a six-and-a-half-hour ses-
______________
83Eric Schmitt and Michael R. Gordon, “British Pressing Partners to Deploy Ground
Troops,” New York Times, May 18, 1999.
84Richard Morin, “Poll Shows Most Americans Want Negotiations on Kosovo,” Wash-
ington Post, May 18, 1999.
85Carla Anne Robbins and Thomas E. Ricks, “Time Is Running Out If Invasion Is to
Remain Option Before Winter,” Wall Street Journal, May 21, 1999.
86Conversation with Air Marshal Sir John Day, RAF, UK Ministry of Defense director of
operations in Allied Force, RAF Innsworth, United Kingdom, July 26, 2000.
48 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
COUNTDOWN TO CAPITULATION
Following an inadvertent attack on a refugee convoy near Djakovica,
Kosovo, on April 14—occasioned in part by a suspected visual
misidentification by the participating USAF F-16 pilots (see Chapter
______________
87Ibid.
88Richard J. Newman, “U.S. Troops Edge Closer to Kosovo,” U.S. News and World Re-
port, June 7, 1999.
89Interview by RAND staff, Washington, D.C., July 11, 2000.
The Air War Unfolds 49
Six)—the altitude floor of 15,000 ft that had been imposed at the start
of the air war was eased somewhat in the southern portion, and
NATO forward air controllers (FACs) flying over Kosovo were cleared
to descend to as low as 5,000 ft if necessary, to ensure positive iden-
tification of ground targets in the KEZ.90 Direct attacks on suspected
VJ positions in Kosovo by B-52s occurred for the first time on May 5
and again the following day. Clark declared afterward that 10 enemy
armor concentrations had been hit and that the Serbs were no longer
able to continue their ethnic cleansing. NATO spokesmen further
reported that enemy troops in the field were running low on fuel and
that VJ and MUP morale had declined.91 A day later, NATO claimed
that it had destroyed 20 percent of the VJ’s artillery and armor de-
ployed in Kosovo. As for infrastructure attacks, only two of the 31
bridges across the Danube in Yugoslavia were said to be still func-
tional by the end of the week. During the second week of May, how-
ever, enemy attack helicopters conducted an attack against the vil-
lage of Kosari along the main supply route for the KLA. They also
served as spotters for VJ artillery against KLA pockets of resistance.92
Those operations indicated that NATO had done an imperfect job of
preventing any and all enemy combat aircraft from flying.
On May 12, roughly 600 Allied Force sorties were launched all told,
including the highest daily number of shooter sorties to date. (See
Figure 3.3 for the overall trend line in U.S. and allied sorties flown
over the 78-day course of Allied Force. Most of the troughs in that
trend line indicate sortie drawdowns or cancellations occasioned by
nonpermissive weather over Serbia.) The multiple waves of succes-
sive large force packages commenced with a sunrise launch of 36 air-
craft, including USAF F-16s and A-10s, RAF Harrier GR. Mk 7s,
French Jaguars and Super Etendards, Italian AMXs, and Canadian
CF-18s. A subsequent late-morning launch featured 32 aircraft,
consisting of RAF Tornado GR. Mk 1s, French Jaguars, and USAF
______________
90The 15,000-ft restriction was never done away with over Serbia and Montenegro,
however, and over Kosovo it was eased only for FACs and for some weapon deliveries
in selected circumstances.
91Hewson, “Allied Force, Part II,” p. 102.
92R. Jeffrey Smith and Dana Priest, “Yugoslavia Near Goals in Kosovo,” Washington
Post, May 11, 1999.
50 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
RAND MR1365-3.3
900
600
Number of sorties
500
400
300
200
100
0
24
31
14
21
28
12
19
26
9
ril
ay
ne
ne
ch
ch
ril
ril
ril
ay
ay
ay
Ap
Ju
Ju
Ap
Ap
Ap
M
ar
ar
M
SOURCE: Hq USAFE/SA.
Three days later, General Jumper declared that NATO had achieved
de facto air superiority over Yugoslavia, enabling attacking aircraft to
“go anywhere we want in the country, any time,” even though the
skies were admittedly “still dangerous.”94 Not long thereafter, an
option became available to attack from the north with 24 F/A-18Ds
of Marine Air Group 31 deployed to Taszar, Hungary. That option
______________
93Hewson, “Allied Force, Part II,” p. 109.
94Ibid., p. 110.
The Air War Unfolds 51
______________
95 Robert Burns, “Use of Apache Copters Is Not Expected Soon,” Philadelphia In-
quirer, May 19, 1999. In what may have been intended as an attempt to lessen the
sting of this leadership ruling, one Army source suggested that sending the Apaches in
had been meant all along merely as a scare tactic to induce Milosevic to negotiate.
The source added that if they had really been intended to be used, the more modern
and capable Apache Longbows would have been deployed instead. “Obviously, it was
just for show, not for go.” Rowan Scarborough, “Apaches Were Sent to Scare Serbs,”
Washington Times, May 21, 1999.
52 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Once Serbia’s air defenses became a less imminent threat, the air war
also saw a heightened use of B-52s, B-1s, and other aircraft carrying
unguided bombs.96 By the end of May, some 4,000 free-fall bombs,
around 30 percent of the total number of munitions expended alto-
gether, had been dropped on known or suspected VJ targets in the
KEZ. There was a momentary resurgence of Serb SAM activity later
that month, with more than 30 SAMs reportedly fired on May 27, the
greatest number launched any night in nearly a month.97 That
heightened activity was assessed as reflecting a determined last-ditch
Serb effort to down at least one more NATO aircraft. (An F-117 had
been shot down during the air war’s fourth night, and a USAF F-16
had later been downed on the night of May 2.)98
______________
96The Block D version of the B-1 employed in Allied Force was configured to carry the
GBU-31 joint direct attack munition (JDAM), but only the B-2 actually delivered that
still-scarce munition.
97It bears noting here that the highly effective GAU-8 30mm cannon carried by the
A-10 saw use only 156 times in Allied Force because of the extreme slant range that
was required by the 5,000-ft altitude restriction (comments on an earlier draft by Hq
USAFE/SA, April 6, 2001). At that range, the principal problem for today’s A-10 pilots
is not hitting the target; it is seeing the target. At a 30-degree dive angle from 5,000 ft,
the slant range to target is 10,000 ft.
98 Glenn Burkins, “Serbs Intensify Effort to Down Allied Warplanes,” Wall Street
Journal, May 28, 1999. In the second instance, the ABCCC drew on instantly accessi-
ble satellite photos and maps maintained in a National Imagery and Mapping Agency
computerized database to identify potential obstacles, such as power lines, in order to
plot a safe course for the rescue helicopter that recovered the downed pilot. Bill Gertz
and Rowan Scarborough, “Inside the Ring,” Washington Times, May 19, 2000. Al-
though there was definitely a pronounced increase in enemy SAM activity during the
night of May 27 in an apparent effort to down a NATO pilot at any cost, it bears
stressing that there were no nights during Allied Force without at least a few SAM
shots, approximately 35 nights with 10 or more shots, and at least 13 nights with 20 or
more shots. The highest number of shots observed (significantly higher than the
number observed on May 27) was on the night of the F-16 loss. Overall, enemy SAM
activity levels tracked closely with allied air attack levels. Low-observable and cruise-
missile-only strikes prompted little enemy IADS reaction, whereas trolling for SAMs
with F-16CJs and CGs and large conventional attack packages always generated a pro-
portionately large enemy reaction. This trend remained consistent throughout the air
war from start to finish. Comments on an earlier draft by Hq USAFE/IN, May 18, 2001.
The Air War Unfolds 53
______________
99Alessandra Stanley, “Albanian Fighters Say They Aid NATO in Spotting Serb Tar-
gets,” New York Times, April 2, 1999.
54 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Throughout most of Allied Force, NATO and the KLA fought parallel
but separate wars against VJ and MUP forces in Kosovo, and both the
U.S. government and the KLA denied coordinating their operations
in advance. NATO did acknowledge, however, that rebel attacks on
the ground had helped flush out VJ troops and armor and to expose
them to allied air strikes on at least a few occasions, and that Clark
had authorized the communication of KLA target location informa-
tion to attacking NATO aircrews indirectly through the ABCCC. The
KLA further acknowledged that NATO air strikes had helped its
ground operations.101 Despite NATO denials throughout the air war
that it was aiding the KLA, it became evident that cooperation be-
tween the two was considerably greater than had been previously
admitted. As reported by KLA soldiers, the KLA had begun as early as
May 10 to supply NATO with target intelligence and other battlefield
information at NATO’s request, with the KLA’s chief of staff, Agim
Ceku, working with NATO officers in northern Albania. While refus-
ing to elaborate on specifics, KLA spokesmen admitted that Ceku had
been the KLA’s principal point of contact with NATO. It was also
Ceku who had participated in Croatia’s 1995 Operation Storm offen-
______________
100 Dana Priest and Peter Finn, “NATO Gives Air Support to Kosovo Guerrillas,”
Washington Post, June 2, 1999.
101 Marjorie Miller, “KLA Vows to Disarm If NATO Occupies Kosovo,” Los Angeles
Times, June 7, 1999.
The Air War Unfolds 55
sive that drove out the Krajina Serbs and helped end the fighting in
Bosnia. 102
Despite this heightened activity in the KEZ during the air war’s final
days, however, the attacks did better at keeping VJ and MUP troops
______________
102 Matthew Kaminski and John Reed, “NATO Link to KLA Rebels May Have Helped
Seal Victory,” Wall Street Journal, July 6, 1999.
103 William Drozdiak and Anne Swardson, “Military, Diplomatic Offensives Bring
About Accord,” Washington Post, June 4, 1999.
104 Tony Capaccio, “JSTARS Led Most Lethal Attacks on Serbs,” Defense Week, July 6,
1999, p. 13.
105 Michael R. Gordon, “A War out of the Night Sky: 10 Hours with a Battle Team,”
New York Times, June 3, 1999.
56 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
dispersed and hidden than they did at actually engaging and killing
them in any significant numbers. Most attack sorties tasked to the
KEZ did not release their weapons against valid military targets, but
rather against so-called dump sites for jettisoning previously unex-
pended munitions, sites that were conveniently billed by NATO tar-
get planners as “assembly areas.” Even the B-52s and B-1s, for all the
free-fall Mk 82 bombs they dropped during the final days, were
tasked with delivering a high volume of munitions without causing
any collateral damage. After the air war ended, it was never estab-
lished that any of the bombs delivered by the B-52s and B-1s had
achieved any militarily significant destructive effects, or that NATO’s
cooperation with the KLA had yielded any results of real operational
value. The steadily escalating attacks against infrastructure targets in
and around Belgrade that were taking place at the same time, how-
ever, were beginning to produce a very different effect on Serb be-
havior.
THE ENDGAME
On June 2, with Operation Allied Force working at peak intensity and
with weather and visibility for NATO aircrews steadily improving,
Russia’s envoy to the Balkans, former Prime Minister Viktor Cher-
nomyrdin, and Finland’s President Martti Ahtisaari, the European
Union representative, flew to Belgrade to offer Milosevic a plan to
bring the conflict to a close. Ahtisaari’s inclusion in the process was
said by one informed observer to have grown out of a suggestion by
Chernomyrdin that value might be gained from including a re-
spected non-NATO player on his mission.106 The same day, after the
two emissaries had essentially served him with an ultimatum that
had been worked out and agreed to previously by the United States,
Russia, the European Union, and Ahtisaari, Milosevic accepted an
international peace proposal. Under the terms of the proposed
agreement, he would accede to NATO’s demands for a withdrawal of
all VJ, MUP, and Serb paramilitary forces from Kosovo; a NATO-led
security force in Kosovo; an unmolested return of the refugees to
______________
106 Comments on an earlier draft by Ivo Daalder of the Brookings Institution, Wash-
ington, D.C., March 15, 2001. Daalder previously served as director for European af-
fairs on the National Security Council staff in 1995 and 1996, where he was responsible
for coordinating U.S. policy for Bosnia.
The Air War Unfolds 57
their homes; and the creation of a self-rule regime for the ethnic Al-
banian majority that acknowledged Yugoslavia’s continued
sovereignty over Kosovo. NATO would continue bombing pending
the implementation of a military-to-military understanding that had
been worked out between NATO and Yugoslavia on the conditions of
Yugoslavia’s force withdrawal. The agreement, which came on the
72nd day of the air effort, was ratified the day after, on June 3, by the
Serb parliament and was rationalized by Milosevic’s Socialist Party of
Serbia on the ground that it meant “peace and a halt to the evil
bombing of our nation.” 107
______________
107 Daniel Williams and Bradley Graham, “Yugoslavs Yield to NATO Terms,” Wash-
ington Post, June 4, 1999.
58 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
NATO refused to commit itself to an early halt to its air attacks, since
its leaders knew that it would be extremely difficult to resume the
bombing once the refugees began coming home. During the negoti-
ations over the terms of Serb withdrawal, however, NATO pilots were
under orders not to attack any enemy positions unless in direct re-
sponse to hostile acts. After the Serb parliament agreed to the cease-
fire, no bombs fell on Belgrade for three consecutive nights. B-52
strikes against dispersed VJ forces, however, continued.
No sooner had this accord been reached in principle than NATO and
Serb military officials failed to reach an understanding on the condi-
tions for VJ and MUP withdrawal. The talks quickly degenerated into
haggling over when NATO would halt its air attacks and whether
Serbia would have more than a week to get its troops out of Kosovo.
The proximate cause of the breakdown in talks was a Serb demand
that the UN Security Council approve an international peacekeeping
force before NATO troops entered Kosovo. That heel-dragging sug-
gested that the Serbs were seeking to soften some of the terms of the
settlement or, perhaps, were looking for more time to continue their
fight with the KLA. Secretary Cohen and General Shelton allowed
that extending the Yugoslav withdrawal by several days would be ac-
ceptable but that they would not countenance any deliberate at-
tempts at delay. More specifically, the implementation of the Serb
withdrawal was hung up on differences over the sequencing of four
events: the start of the enemy pullout, a pause in NATO bombing,
the passage of a UN resolution, and the entry of international peace-
keepers with a “substantial NATO content.” In response to this will-
ful foot-dragging, NATO’s attacks, which initially had been scaled
back after Milosevic accepted the proposed peace plan, resumed
their previous level of intensity.
On June 7, at the same time as the talks were under way, VJ forces
launched a renewed counterattack against the KLA in an area south
of Mount Pastrik, where the two sides had been locked in an artillery
duel since May 26. For a time, a major breakthrough in NATO’s air
effort was thought to have occurred when the defending KLA forces
flushed out VJ troops who had been dispersed around Mount Pastrik,
creating what NATO characterized as a casebook target-rich envi-
The Air War Unfolds 59
Whatever the case, the following day the United States, Russia, and
six other member-states agreed on a draft UN Security Council reso-
lution to end the conflict. The resolution called for a complete with-
drawal of Serb troops, police, and paramilitary forces from Kosovo
and for all countries to cooperate with the war crimes tribunal that
had indicted Milosevic. 110 The sequence finally agreed to was that
the Serb force withdrawal would commence, NATO would concur-
rently halt its bombing, and only after those two actions occurred
would the Security Council vote on the text of the agreement. The
last provision was a token concession to Russia and China, whose
representatives had insisted that the bombing be stopped before any
Security Council vote was taken.
______________
108 R. Jeffrey Smith and Molly Moore, “Plan for Kosovo Pullout Signed,” Washington
Post, June 10, 1999.
109 William Drozdiak, “Yugoslav Troops Devastated by Attack,” Washington Post, June
9, 1999.
110 Smith and Moore, “Plan for Kosovo Pullout Signed.”
111 At one point in the negotiations, the VJ military delegation leader, Colonel General
Svetozar Marjanovic, abruptly walked out of the talks, stating that he needed to
60 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
NATO finally stopped the bombing upon verifying that the Serb
withdrawal had begun, after which the UN Security Council ap-
proved, by a 14-0 vote with China abstaining, a resolution putting
Kosovo under international civilian control and the peacekeeping
force under UN authority. With that, President Clinton declared that
NATO had “achieved a victory.”112
_____________________________________________________________
“consult with authorities in Belgrade.” He made it only to a border post and returned
to the negotiating table within an hour.
112 Tim Weiner, “From President, Victory Speech and a Warning,” New York Times,
June 11, 1999.
113 See Julian Barger, “Bloody Paper Chain May Link Torture to Milosevic,” The
Guardian, June 18, 1999.
114 Ian Black and John Hooper, “Serb Savagery Exposed,” The Guardian, June 18, 1999.
The Air War Unfolds 61
informally for 10 days when the first Serb troops began leaving
Kosovo. The departure of the last Serb forces and the arrival of the
KFOR peacekeepers effectively brought an end to Yugoslav control
over a province that had been a special and even sacred preserve of
Serbia for centuries.
Initial estimates just before the cease-fire went into effect claimed
that the air war had taken out 9 percent of Serbia’s soldiers (10,000 of
114,000), 42 percent of its aircraft (more than 100 of 240), 25 percent
of its armored fighting vehicles (203 of 825), 22 percent of its artillery
pieces (314 of 1,400), and 9 percent of its tanks (120 of 1,270).115
After the cease-fire, the Pentagon claimed that the operation had de-
stroyed 450 enemy artillery pieces, 220 armored personnel carriers,
120 tanks, more than half of Yugoslavia’s military industry, and 35
percent of its electrical power-generating capacity.116 General Shel-
ton reported that 60 percent of the infrastructure of the Yugoslav 3rd
Army, the main occupying force in Kosovo, had been destroyed,
along with 35 percent of the 1st Army’s infrastructure and 20 percent
of the 2nd Army’s.117 The U.S. Air Force’s deputy chief of staff for air
and space operations, Lieutenant General Marvin Esmond, an-
nounced that the allied bombing effort had destroyed a presumed 80
percent of Yugoslavia’s fixed-wing air force, zeroed out its oil refining
capability, and eliminated 40 percent of its army’s fuel inventory and
40 percent of its ability to produce ammunition. Many of these ini-
tial assessments were later discovered to have been overdrawn by a
considerable margin.
In the final tally, allied aircrews flew 38,004 out of a planned 45,935
sorties in all, of which 10,484 out of a planned 14,112 were strike sor-
ties.118 A later report to Congress by Secretary Cohen and General
Shelton claimed that more than 23,300 combat missions, including
______________
115 Michael R. Gordon and Eric Schmitt, “Shift in Targets Let NATO Jets Tip the Bal-
ance,” New York Times, June 5, 1999.
116 Weiner, “From President, Victory Speech and a Warning.”
117 Bradley Graham, “Air Power ‘Effective, Successful,’ Cohen Says,” Washington Post,
June 11, 1999.
118 Operation Allied Force and Operation Joint Guardian briefing charts dated August
19, 1999, provided to the author by Air Marshal Sir John Day, RAF, UK Ministry of De-
fense director of operations in Allied Force, RAF Innsworth, United Kingdom, July 25,
2000.
62 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
119 Secretary of Defense William S. Cohen and Chairman of the Joint Chiefs of Staff
General Henry H. Shelton, Kosovo/Operation Allied Force After-Action Report,
Washington, D.C., Department of Defense, Report to Congress, January 31, 2000, p. 87.
120 Ibid., p. 68.
121 Operation Allied Force and Operation Joint Guardian briefing charts dated August
19, 1999.
The Air War Unfolds 63
RAND MR1365-3.4
Total: 30,018
UAVs
(496)
Fighter
(8,889)
Intratheater
airlift
(11,480)
Bomber
(322)
Special Tanker
Ops (6,959)
ISR
(834) (1,038)
______________
122 “AWOS Fact Sheet,” Hq USAFE/SA, December 17, 1999. See also William M. Arkin,
“Top Air Force Leaders to Get Briefed on Serbia Air War Report,” Defense Daily, June
13, 2000, p. 1. As attested by cockpit display videotapes released to the press
throughout the air war, allied air attacks succeeded in taking out quite a few more SA-6
launchers than those accounted for here. However, since the STRAIGHT FLUSH radar
formed the core of an SA-6 battery, the battery was considered operational until the
STRAIGHT FLUSH was destroyed. Comments on an earlier draft by Hq USAFE/IN,
May 18, 2001.
64 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
U.S.
500 Non-U.S.
400
Number of munitions
300
200
100
0
24
31
14
21
28
12
19
26
9
ril
ay
ne
ne
h
ril
ril
ril
ay
ay
ay
Ap
M
c
Ju
Ju
Ap
Ap
Ap
M
ar
ar
M
SOURCE: Hq USAFE/SA.
RAND MR1365-3.6
1600
Non-PGM
1400 PGM
1200
Number of munitions
1000
800
600
400
200
0
24
31
14
21
28
12
19
26
9
ril
ay
ne
ne
ch
ch
ril
ril
ril
ay
ay
ay
Ap
Ju
Ju
Ap
Ap
Ap
M
ar
ar
M
SOURCE: Hq USAFE/SA.
RAND MR1365-3.7
Total: 28,018
Allied
(4,703)
17%
83%
U.S.
(23,315)
Safe recovery. An F-117 stealth attack aircraft lands at Aviano Air Base, Italy, just after sunrise fol-
lowing a night mission into the most heavily defended portions of Serbia. During the air war’s
fourth night, an F-117 was downed just northwest of Belgrade, most likely by a lucky SA-3 shot, in
the first-ever loss of a stealth aircraft in combat. (The pilot was promptly retrieved by CSAR forces.)
Heavy players. A venerable USAF B-52H bomber stands parked on the ramp at RAF Fairford,
England, as a successor-generation B-1B takes off on a mission to deliver as many as 80 500-lb Mk
82 bombs or 30 CBU-87 cluster bomb units against enemy barracks and other area targets. An
AGM-86C CALCM fired from standoff range by a B-52 was the first allied weapon to be launched in
the war.
Final checks. Two Block 40 F-16CGs from the 555th Fighter Squadron at Aviano taxi into the arm-
ing area just short of the runway for one last look by maintenance technicians before taking off on
a day mission to drop 500-lb GBU-12 laser-guided bombs on “flex” targets of opportunity in Serbia
or Kosovo, as directed by airborne FACs and as approved, in some cases, by the CAOC.
Burner takeoff. An F-15E from the 494th Fighter Squadron home-based at RAF Lakenheath,
England, clears the runway at Aviano in full afterburner, with CBU-87 cluster munitions shown
mounted on its aftmost semiconformal fuselage weapons stations. Eventually, some F-15E strike
sorties into Serbia and Kosovo were flown nonstop to target and back directly from Lakenheath.
On the cat. A U.S. Navy F/A-18C assigned to Fighter/Attack Squadron 15 is readied for a catapult
launch for an Allied Force day combat mission from the aircraft carrier USS Theodore Roosevelt
cruising on station in the Adriatic Sea. On April 15, carrier-based F/A-18s figured prominently in a
major CAOC-directed air strike on the Serb air base at Podgorica, Montenegro.
SAM hunter. This Block 50 F-16CJ in the arming area at Aviano shows an AGM-88 high-speed anti-
radiation missile (HARM) mounted on the left intermediate wing weapons station, with an AIM-
9M air-to-air missile on the outboard station and an AIM-120 AMRAAM on the wingtip missile rail.
The USAF’s F-16CJ inventory was stressed to the limit to meet the SEAD demand of Allied Force.
Combat support. U.S. Navy and Marine Corps EA-6B Prowler electronic warfare aircraft, like this
one shown taxiing for takeoff at Aviano, provided extensive and indispensable standoff jamming of
enemy early warning and IADS fire-control radars to help ensure unmolested allied strike opera-
tions, including B-2 and F-117 stealth operations, against the most heavily defended enemy targets
in Serbia.
Task Force Hawk. A U.S. Army AH-64 Apache attack helicopter flares for landing at the Rinas air-
port near Tirana, Albania, following a ferry flight from its home base at Illesheim, Germany. In all,
24 Apaches were dispatched to Albania with the intent to be used in Operation Allied Force, but
none saw combat in the end because of concerns for the aircraft’s prospects for survival in hostile
airspace.
Cramped spaces. This USAF C-17 parked on the narrow ramp at Rinas airport, incapable of
accommodating the larger C-5, was one of many such aircraft which provided dedicated mobility
service to TF Hawk. In more than 500 direct-delivery lift sorties altogether, C-17s moved 200,000-
plus short tons of equipment and supplies to support the Army’s deployment within the span of
just a month.
Eagle eye. Ground crewmen at RAF Lakenheath prepare a LANTIRN targeting pod to be mounted
on an F-15E multirole, all-weather fighter. The pod, also carried by the Navy F-14D and the USAF
F-16CG, contains a forward-looking infrared (FLIR) sensor for target identification at standoff ranges
day or night, as well as a self-contained laser designator for enabling precision delivery of LGBs.
Deadly force. Munitions technicians at RAF Fairford prepare a CBU-87 cluster bomb unit for load-
ing into a USAF B-1B bomber in preparation for a mission against fielded Serbian forces operating
in Kosovo. With a loadout of 30 CBU-87s—more than five times the payload of an
F-15E—the B-1 can fly at fighter-equivalent speeds more than 4,200 nautical miles unrefueled.
Help from an ally. One of 18 CF-18 Hornet multirole fighters deployed in support of Allied Force
from Canadian Forces Base Cold Lake, Alberta, Canada, is parked in front of a hardened aircraft
shelter at Aviano. The aircraft mounts two 500-lb GBU-12 laser-guided bombs on the outboard
wing pylons and two AIM-9M air-to-air missiles on the wingtip rails.
Force protection. A security guard stands watch over a USAF C-17 airlifter at the Rinas airport in
support of the U.S. Army’s Apache attack helicopter deployment to Albania. For a time, reported
differences between on-scene Air Force and Army commanders with respect to who was ultimate-
ly responsible for the airfield made for discomfiting friction within the U.S. contingent.
Flexing into the KEZ. An AGM-65 Maverick-equipped A-10 from the USAF’s 52nd Fighter Wing
stationed at Spangdahlem Air Base, Germany, takes off from Gioia del Colle to provide an on-call
capability against possible Serb targets detected in Kosovo by allied sensors, including the TPQ-36
and TPQ-37 firefinder radars operated by the U.S. Army on the high ground above Tirana, Albania.
Round-the-clock operations. An F-16 pilot readies himself for a night mission over Serbia, his hel-
met shown fitted with a mount for night-vision goggles. Used in conjunction with compatible
cockpit lighting, NVGs made possible night tactics applications, including multiaircraft formations
and simultaneous bomb deliveries, which otherwise could only have been conducted during day-
light.
Night refueling. A USAF F-15C air combat fighter, shown here through a night-vision lens, moves
into the precontact position to take on fuel from a KC-135 tanker before resuming its station to
provide offensive counterair protection for attacking NATO strikers. With a loss of six MiGs in aer-
ial combat encounters the first week, Serb fighters rarely rose thereafter to challenge NATO’s con-
trol of the air.
Splash one Fulcrum. A team of U.S. military personnel examines the remains of an enemy MiG-
29 fighter (NATO code name Fulcrum) which was shot down in Bosnian airspace by a USAF F-15C
on the afternoon of March 26, 1999. The downed aircraft, which appeared to have strayed from its
planned course due to a loss of situation awareness by its pilot, brought to five the number of
MiG-29s destroyed in early Allied Force air encounters.
Hard-target killer. This F-15E pilot, a USAF major assigned to the 494th Fighter Squadron, looks
over a 4,700-lb electro-optically-guided GBU-28 bunker-buster munition mounted on his aircraft’s
centerline stores station. The aircraft, one of a two-ship flight of F-15Es (call sign Lance 31 and 32),
delivered the weapon on April 28, 1999, against an underground hangar at the Serb air base at
Podgorica.
Precision attack. In April 1999, a single B-2 achieved six accurately placed GBU-31 JDAM hits
against six runway-taxiway intersections at the Obvra military airfield in Serbia, precluding opera-
tions by enemy fighters until repairs could be completed. This post-strike image graphically shows
the B-2’s ability with JDAM to achieve the effects of mass without having to mass, regardless of
weather.
A bridge no more. Another post-strike battle-damage assessment image shows this bridge in
Serbia cut in two places by a precision bombing attack. Sometimes enemy bridges were dropped
at the behest of NATO target planners to prevent the flow of traffic over them. At other times, they
were attacked and damaged to sever key fiber-optic communications lines that were known to run
through them.
Before and after. This bridge over the Danube River near Novi Sad in Serbia, shown here in both
pre- and post-strike imagery, was all but completely demolished by precision bombing on Day 9 of
Allied Force as Phase III of the air war, for the first time, ramped up operations to include attacks
against not only Serbian IADS and fielded military assets but also key infrastructure targets.
Effects-based targeting. For three consecutive nights beginning on May 24, U.S. aircraft struck
electrical power facilities in Belgrade, Novi Sad, and Nis, the three largest cities in Serbia, shutting
off electrical power to 80 percent of Serbia. This transformer yard in Belgrade was one such target
that was attacked in what was arguably the most influential strike of Allied Force to that point.
Chapter Four
WHY MILOSEVIC GAVE UP WHEN HE DID
67
68 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
We may never know for sure what mix of pressures and inducements
ultimately led Milosevic to admit defeat, at least until key Serb
archival materials become available or those closest to Milosevic
during the air war become disposed to offer first-hand testimony.
Asked by a reporter why Milosevic folded if the bombing had not
defeated him militarily, Clark, who knew the Serb dictator well from
previous negotiating encounters, replied: “You’ll have to ask Milo-
sevic, and he’ll never tell you.”3 Yet why Milosevic gave in and why
he did so when he did are by far the most important questions about
the air war experience, since the answers, insofar as they are know-
able, will help to lay bare the coercive dynamic that ultimately swung
the outcome of Allied Force. It need hardly be said that such insight
can be of tremendous value in informing any strategy ultimately
chosen by the United States and its allies for future interventions of
that sort. Accordingly, it behooves analysts to make every effort to
delve further into this innermost mystery of the air war, since even
approximate answers, if buttressed by valid evidence, are almost
certain to be more useful to senior policymakers than most “lessons”
of a more technical nature regarding how specific systems worked
and how various procedural aspects of the operation could have
been handled better, important as the latter questions are.
In the search to understand what ultimately occasioned NATO’s
success, one can, of course, insist that air power alone was the cause
of Milosevic’s capitulation in the tautological sense that Allied Force
was an air-only operation and that in its absence, there would have
been no reason to believe that he would have acceded to NATO’s
demands.4 Yet as crucial as the 78-day bombing effort was in
bringing Milosevic to heel, there is ample reason to be wary of any
intimation that NATO’s use of air power produced that ending with-
out any significant contribution by other factors. On the contrary,
numerous considerations in addition to the direct effects of the
bombing in all likelihood interacted to produce the Serb dictator’s
eventual decision to cave in. As Ivo Daalder and Michael O’Hanlon
have remarked, in a balanced reflection on this point, “air power
______________
3 Michael Ignatieff, “The Virtual Commander: How NATO Invented a New Kind of
War,” The New Yorker, August 2, 1999, p. 31.
4 See, for example, Rebecca Grant, “Air Power Made It Work,” Air Force Magazine,
November 1999, pp. 30–37.
Why Milosevic Gave Up When He Did 69
______________
5Ivo H. Daalder and Michael E. O’Hanlon, Winning Ugly: NATO’s War to Save Kosovo,
Washington, D.C., The Brookings Institution, 2000, p. 184.
6This official, in an interview with RAND staff members in Washington on June 11,
2000, further claimed that the White House was not surprised when Milosevic
accepted the deal on June 3, since the administration was confident that once Cher-
nomyrdin had agreed to NATO’s terms, it was merely a matter of time before a suc-
cessful denouement would be reached, considering that Chernomyrdin knew Milose-
vic’s bottom line and would not have signed up for any arrangement that he knew
Milosevic would not accept. What was surprising, the official said, was that Milosevic
did not first seek to buy time by proffering more “half-loaf” compromise deals.
70 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
the Kosovo standoff, he knew that he had lost any remaining trace of
international backing.
On top of that was the sense of walls closing in that Milosevic must
have had when he was indicted as a war criminal by a UN tribunal
only a week before his loss of Moscow’s support. On May 27, that
tribunal charged Milosevic and four of his senior aides—including
General Dragoljub Ojdanic, the Yugoslav army chief, and Vlajko Sto-
jilkovic, the interior minister responsible for the MUP—with crimes
against humanity for having deported more than 700,000 ethnic Al-
banians and having allegedly murdered 340 innocents, mostly young
men. Even if that indictment did not give Milosevic pause in and of
itself, it almost surely closed the door on any remaining chance that
Russia might change course and resume its support for him.
Yet a third factor, this one a direct second-order result of the bomb-
ing, may have been mounting elite pressure behind the scenes. As
the air attacks encroached more on Belgrade proper, Secretary Co-
hen reported that senior VJ leaders had begun sending their families
out of Yugoslavia, following a similar action earlier by members of
the Yugoslav political elite and reflecting possible concern among
top-echelon commanders that Milosevic had led them down a blind
alley in choosing to take on the United States and NATO.7 U.S. offi-
cials indicated that during the last week of the air war, VJ leaders had
swung from supporting Milosevic on Kosovo to openly rebelling and
pressuring the Serb dictator to agree to NATO’s terms. Cohen’s re-
port of increasing demoralization among the VJ’s most senior leaders
as they helplessly watched the escalating destruction all around
them gave rise to hopes within the Clinton administration that Milo-
sevic might be looking for a face-saving way out. 8 The fact that the
bombing effort caused more infrastructure damage during its last
week than during its entire first two months was thought by some to
have reawakened old tensions between Milosevic and an army lead-
ership that was said to have never fully trusted him.
______________
7Daniel Williams and Bradley Graham, “Milosevic Admits to Losses of Personnel,”
Washington Post, May 13, 1999.
8 Interview with Secretary of Defense William S. Cohen, “Milosevic Is Far Weaker
Now,” USA Today, May 14, 1999.
Why Milosevic Gave Up When He Did 71
______________
9Paul Richter, “Officials Say NATO Pounded Milosevic into Submission,” Los Angeles
Times, June 5, 1999. The possible effects of the bombing on what one might call sec-
ond-tier Serb leaders are especially noteworthy, in that they suggest that the elite sub-
structure of an enemy’s hierarchy may make for more lucrative leadership targets than
the “big guys.” Unlike the topmost political leaders, these second-tier individuals
have “retirement plans,” in that they have options to recoup their interests under a
new regime. They thus may be more malleable than their bosses, even as they are of-
ten critical to their bosses’ survival. I am grateful to Colonel Robert Owen, USAF, for
having suggested this intriguing idea to me.
10Doyle McManus, “Clinton’s Massive Ground Invasion That Almost Was,” Los Ange-
les Times, June 9, 2000. A persistent concern that tended to inhibit a truly aggressive
use of such information entailed the liability implications of information attacks
against foreign bank accounts, as well as official worries about the Pandora’s box that
might be opened if the United States began playing that game, thus rendering its own
economy susceptible to similar measures in return.
11 Tom Walker, “Bomb Video Took Fight out of Milosevic,” London Sunday Times,
January 30, 2000.
72 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
fighting vehicle (AFV). This time, only three days before Milosevic fi-
nally called it quits, Washington gave Clark permission to send in
another engineering battalion to make the road capable of support-
ing M1A2 Abrams tanks and artillery. 12
Beyond that, Milosevic may have gotten wind of a secret NATO plan
for a massive ground invasion code-named Plan B-minus, which was
slated to be launched the first week of September if approved by
NATO’s political leaders. In support of this plan, Britain had agreed
to contribute the largest single national component up to that time
(50,000 troops) to an envisaged 170,000-man contingent; the United
States would have contributed at least 100,000 more. Developed by a
secret planning team at NATO’s military headquarters in Mons, Bel-
gium, Plan B-minus relied heavily on previous plans going back to
June 12, 1998, which featured six land-attack options, including a full
invasion of Serbia itself (Plan Bravo, with 300,000 NATO troops). The
chief of Britain’s defense staff, General Sir Charles Guthrie, later con-
firmed the outlines of this plan. 13 Milosevic was said by a well-
placed NATO source to have been at least broadly informed of NATO
thinking with respect to it. Indeed, as the UK Ministry of Defense’s
director of operations in Allied Force, RAF Air Marshal Sir John Day,
later commented, “the decision to increase KFOR was militarily right
in itself, but it was also a form of heavy breathing on Milosevic and a
subtle way of moving to B-minus while keeping the coalition to-
gether. The move also had the effect of shortening our timelines for
B-minus. It is true that the forces that were being prepared for
KFOR-plus were the core elements of what would then have become
B-minus, the full ground invasion.”14
______________
12Dana Priest, “A Decisive Battle That Never Was,” Washington Post, September 19,
1999.
13 Patrick Wintour and Peter Beaumont, “Revealed: The Secret Plan to Invade
Kosovo,” London Sunday Observer, July 18, 1999.
14Peter Beaumont and Patrick Wintour, “Leaks in NATO—and Plan Bravo Minus,”
London Sunday Observer, July 18, 1999.
74 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
15Michael R. Gordon, “NATO Says Serbs, Fearing Land War, Dig In on Border,” New
York Times, May 19, 1999.
16Jane Perlez, “Clinton and the Joint Chiefs to Discuss Ground Invasion,” New York
Times, June 2, 1999.
17For details, see Steven Erlanger, “NATO Was Closer to Ground War in Kosovo Than
Is Widely Realized,” New York Times, November 7, 1999.
18McManus, “Clinton’s Massive Ground Invasion That Almost Was.” NATO com-
manders were asking for three months to assemble the invasion force.
Why Milosevic Gave Up When He Did 75
forced entry on the ground if the bombing did not produce an ac-
ceptable settlement soon.
Some, however, have made more of this sequence of events than the
evidence warrants. In the early wake of the successful conclusion of
Operation Allied Force, revisionist claims began emanating from
some quarters suggesting that the air effort had been totally ineffec-
tive and that, in the end, it had been Milosevic’s fear of a NATO
ground invasion that induced him to capitulate.19 Clark himself, in
his memoirs, indicated his belief that by mid-May, NATO “had gone
about as far as possible with the air strikes” and that in the end, it
had been the Apache deployment and the prospect of a NATO
ground intervention that, “in particular, pushed Milosevic to con-
cede.”20 That notwithstanding the all-but-conclusive evidence Clark
presented elsewhere throughout his book that NATO’s top political
leaders were nowhere near having settled on a definitive invasion
plan—let alone decided to proceed with such a plan should the
bombing prove unavailing.21 Even viewed in the most favorable light
conceivable, such far-reaching claims on behalf of the implied
ground threat defy believability because any NATO land invasion,
however possible it may eventually have been, would have taken
months, at a minimum, to prepare for and successfully mount.
______________
19A recent example of this countercontention dismissed the claims of unspecified “air
power enthusiasts” and posited instead that “the decision to commit ground forces [a
decision which, in fact, had not been made at the time of Milosevic’s capitulation] was
critical to NATO’s success.” Brigadier General Huba Wass de Czege, USA (Ret.) and
Lieutenant Colonel Antulio J. Echevarria II, USA, “Precision Decisions: To Build a
Balanced Force, the QDR Might Consider These Four Propositions,” Armed Forces
Journal International, October 2000, p. 54.
20General Wesley K. Clark, Waging Modern War: Bosnia, Kosovo, and the Future of
Combat, New York, Public Affairs, 2001, pp. 305, 425, emphasis added.
21The most compelling of such evidence cited by Clark was the May 28 statement by
Secretary of Defense Cohen, made less than a week before Milosevic capitulated, that
“there is no consensus for a ground force. And until there is a consensus, we should
not undertake any action for which we could not measure up in the way of perfor-
mance. . . . And so, there is a very serious question in terms of trying to push for a
consensus that you really diffuse or in any way diminish the commitment to the air
campaign. The one thing we have to continue is to make sure we have the allies
consolidated in strong support of the air campaign. They are. And they are in favor of
its intensification. So that’s where we intend to put the emphasis.” Ibid, p. 332.
76 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
There is no question that by the end of May, NATO had yielded to the
inevitable and embraced in principle the need for a ground invasion
should the bombing continue to prove indecisive. There also is every
reason to believe that awareness of that change in NATO’s position
on Milosevic’s part figured importantly in his eventual decision to
capitulate. There is no basis, however, for concluding that the mere
threat of a land invasion somehow overshadowed the continuing,
here-and-now reality of NATO’s air attacks as the preeminent con-
sideration accounting for that decision. There also is little benefit to
be gained from the misguided efforts by air and land power partisans
alike to argue the relative impact of the air attacks and ground threat
in simplistic either-or terms. It detracts not in the least from the air
war’s signal accomplishments to concede that developments on the
land-invasion front almost surely were part of the chemistry of Milo-
sevic’s concession decision. Although any impending ground inter-
vention was months away at best, there is no question that both the
Clinton administration and the principal NATO allies had made up
______________
22However, by dispersing their assets and selectively emitting with their radars, Serb
IADS operators forced NATO aircrews to remain wary to the very end and denied them
the freedom to operate at will in hostile airspace. Although the Serbs’ repeated at-
tempts to bring down NATO aircraft frequently came in the form of ineffective ballistic
launches, the launches were amply disconcerting to allied pilots, who were forced to
threat-react—often aggressively—to ensure their own safety. Many guided shots in
accordance with IADS doctrine were also fired against attacking allied aircraft, requir-
ing even more aggressive and hair-raising countertactics by the targeted aircraft. A
first-hand account of one such episode is reported in Dave Moniz, “Eye-to-Eye with a
New Kind of War,” Christian Science Monitor, March 23, 2000.
23Quoted in Tyler Marshall and Richard Boudreaux, “Crisis in Yugoslavia: How an
Uneasy Alliance Prevailed,” Los Angeles Times, June 6, 1999.
Why Milosevic Gave Up When He Did 77
their minds on the need to do something along those lines should the
air war continue to prove unavailing. In light of that, as two RAND
colleagues have suggested, “in assessing NATO air attacks on Serbia,
analysts should focus not on the role air power played instead of a
ground invasion . . . but on the role it played in combination with the
possibility of one.”24
Yet to those on the operation’s receiving end far removed from such
concerns, it must have seemed, certainly by the end of the second
month, as though NATO was prepared to keep escalating and to
continue bombing indefinitely. From Milosevic’s viewpoint, new
targets were being attacked with mounting regularity after the NATO
summit of April 23–25, and ever more infrastructure targets were be-
ing hit with seemingly no end in sight. Moreover, one might surmise
that even the inadvertent Chinese embassy bombing played an indi-
rect part in inducing Milosevic to capitulate. Whatever U.S. and
NATO officials said about that incident for the public record, Milo-
sevic may have thought that the bombing had been intentional and
that it presaged both a lifting of NATO’s target limitations and worse
damage yet to come. As if to affirm that fear after the fact, USAFE’s
commander, General John Jumper, later disclosed that with the in-
creased number of strike aircraft that had become available in the-
______________
24Daniel L. Byman and Matthew C. Waxman, “Kosovo and the Great Air Power De-
bate,” International Security, Spring 2000, p. 15.
78 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
ater by late May, the operation’s intent was to employ FACs and be-
gin attacking kill boxes all throughout Serbia, not just in Kosovo, and
to go at will after tunnels, bridges, storage areas, and other military
targets of interest.25
______________
25General John Jumper, USAF, “Oral Histories Accomplished in Conjunction with
Operation Allied Force/Noble Anvil.”
26Stephen T. Hosmer, The Conflict over Kosovo: Why Milosevic Decided to Settle When
He Did, Santa Monica, California, RAND, MR-1351-AF, 2001.
27The latter of these two concerns was more an issue for Milosevic than the former.
Had he been seriously worried about a NATO presence that might actually encroach
into Serbia, as opposed to just taking effective control of Kosovo (his real fear),
he would have sought to head off that possibility at Rambouillet. He never did. I
am grateful to Ivo Daalder of the Brookings Institution for bringing this point to my
attention.
Why Milosevic Gave Up When He Did 79
Where Milosevic blundered even more grievously than did NATO (in
the latter’s faulty assumption that just a few days of bombing would
suffice) was in unleashing the full brunt of his ethnic cleansing cam-
paign almost immediately after Allied Force began. No doubt he cal-
culated that Operation Horseshoe would quickly empty Kosovo of its
ethnic Albanian populace and thus enable him to move directly
against the KLA, eliminate it as a continued factor affecting any ulti-
mate political outcome, and, along the way, solve his ethnic problem
in Kosovo with a fait accompli. Alternatively, or perhaps in addition,
he may also have been trying to signal his own determination to
NATO, although there is no “smoking-gun” evidence to this effect.
After all, the main lesson he likely drew from Deliberate Force in 1995
was that he gave up the fight just a few days too early. Most assess-
ments of Deliberate Force include arguments that NATO was ap-
proaching the end of its rope politically and militarily because of a
______________
28 Justin Brown, “Why U.S. Bombs Failed to Topple Milosevic,” Christian Science
Monitor, March 24, 2000.
80 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
29See, in particular, Colonel Robert Owen, USAF, ed., Deliberate Force: A Case Study
in Effective Air Campaigning, Maxwell AFB, Alabama, Air University Press, January
2000, pp. 455–522.
30Stephen Hosmer has pointed out that the ethnic cleansing hardened NATO’s re-
solve in another way as well: Only a NATO military presence in Kosovo would have
convinced the refugees to go back to their homes, and no outcome short of the latter
would have been acceptable to NATO.
Why Milosevic Gave Up When He Did 81
In the end, however inefficient the air war may have been because of
its need to honor U.S. and NATO domestic political realities, the
manner in which it was conducted (avoiding friendly fatalities and
minimizing noncombatant enemy casualties) nevertheless effec-
tively countered and ultimately neutralized Milosevic’s strategy by
keeping NATO’s cohesion intact to the very end. In response, the
Serb dictator most likely opted to accept NATO’s demands simply
out of a rational calculation that he had nothing to gain and much to
risk by holding out any longer. Indeed, as the endgame neared, one
can imagine how he may even have begun to harbor dark visions of
being gunned down in the street, in the grim manner of the Ceauces-
______________
31Barry R. Posen, “The War for Kosovo: Serbia’s Political-Military Strategy,” Interna-
tional Security, Spring 2000, p. 75. One can, however, question Posen’s subsequent
suggestion that Milosevic achieved “some political success” by holding out as long as
he did, considering that he lost control of Kosovo, suffered heavy damage to his infra-
structure and economy, and ultimately was defeated in a fair election, arrested, and
jailed for having committed crimes against the state.
82 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
cus after their control over Romania collapsed in 1991. Said a source
close to the Yugoslav government: “I can’t pinpoint an exact
moment when Milosevic finally listened, but there was tremendous
pressure from all sides; the West, his inner circle, and his wife. It was
building up, and eventually he just let go.”32
______________
32“NATO’s Game of Chicken,” Newsweek, July 26, 1999, p. 59.
33For detailed amplification on this point, see Hosmer, The Conflict over Kosovo: Why
Milosevic Decided to Settle When He Did.
34 Quoted in John T. Correll, “Lessons Drawn and Quartered,” Air Force Magazine,
December 1999, p. 2.
Why Milosevic Gave Up When He Did 83
The precise and measured nature of the attacks that were being con-
ducted against leadership and infrastructure targets in the heart of
the Yugoslav capital on a daily basis only became fully apparent to
outside observers after they had a chance to inspect the results up
close. As one American reporter who visited Belgrade after the war
remarked tellingly: “Like ice-pick punctures in the neck, the chilling
quality of the strikes was not their size but their placement. We
stopped at an intersection in the heart of the city. At each corner of
the intersection, but only at each corner, there were ruins. The Ser-
bian government center, the foreign ministry and two defense min-
istry buildings had been reduced to rubble or were fire-gutted shells.
The precision of the destruction suggested a war with an invisible,
all-seeing enemy and a city helpless to protect itself.”36
______________
35Quoted in Robert Hewson, “Allied Force, Part II: Overwhelming Air Power,” World
Air Power Journal, Winter 1999/2000, p. 97. Three days later, Draskovic was fired by
Prime Minister Momir Bulatovic for having made that statement.
36Blaine Harden, “The Milosevic Generation,” New York Times Magazine, August 29,
1999, p. 34.
37“Sacked Yugoslav Air Chief Killed,” London Times, June 2, 1999. See also William
Drozdiak and Steven Mufson, “NATO Sending Tough Terms to Belgrade,” Washington
Post, June 2, 1999.
84 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
NATO not only would not relent, but also was determined to prevail
and had both the technical and political wherewithal to do so. Sec-
ond, given the incapacity of the Serb IADS to shoot down significant
numbers of allied aircraft, it further convinced him that his own de-
feat sooner or later was inevitable. Although its resolve was slow in
coming, NATO finally showed that it would not be moved by the
public outcry over collateral damage and could sustain the bombing
indefinitely, at a negligible cost in terms of friendly losses. As with
Iraq’s forces during Operation Desert Storm, the VJ’s leaders, no less
than Milosevic, must have found NATO’s ability to inflict unrelenting
damage on their country with virtual impunity to be profoundly de-
moralizing. Before June 3, the commander of the VJ’s 3rd Army in
Kosovo, General Nebojsa Pavkovic, had argued that his forces re-
mained more or less intact and that they could defend Serbia if put to
the test. After Ahtisaari and Chernomyrdin delivered NATO’s ultima-
tum on June 2 and a cease-fire was agreed to, however, he reportedly
declared to a group of disconcerted VJ reservists that Serbia’s leaders
had been put on notice by the Russians that if NATO’s terms were
rejected, “every city in Serbia would be razed to the ground. The
bridges in Belgrade would be destroyed. The crops would all be
burned. Everyone would die.”38
______________
38 Quoted in Chris Hedges, “Angry Serbs Hear a New Explanation: It’s All Russia’s
Fault,” New York Times, July 16, 1999.
Why Milosevic Gave Up When He Did 85
______________
39Indeed, from a low of fewer than 100 daily strike sorties flown during the air war’s
fifth night, the bombing effort intensified steadily and uninterruptedly to almost three
times that number by the eve of Milosevic’s capitulation on June 3. Briefing by the
chairman of the Joint Chiefs of Staff, June 10, 1999, cited in Major General Eitan Ben-
Eliahu, commander, Israeli Air Force, “Air Power in the 21st Century: The Impact of
Precision Weapons,” Military Technology, April 2000, p. 40.
40It bears acknowledging here, however, that only the authoritative report of NATO’s
intent to proceed with an eventual ground invasion, should the bombing alone fail to
dislodge Milosevic, finally convinced Moscow to play its constructive role in June
1999. Russia’s deploying of Chernomyrdin helped negotiate an international military
presence in Kosovo, thus warding off a NATO-only presence and preserving at least
some Russian influence in the Balkans. On this point, see the informed comment of-
fered by former Russian foreign ministry Balkan official Oleg Levitin, “Inside
Moscow’s Kosovo Muddle,” Survival, Spring 2000, p. 138.
86 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
41William Arkin, “Yugoslavia Trip Report,” September 8, 1999. In a similar vein, Karl
Mueller suggested that “while it was not clear how NATO was going to win, it certainly
would continue the effort until it managed to do so. From this perspective, it was
not what NATO was bombing that mattered, but the fact that it was continuing to
bomb. . . .” Karl Mueller, “Deus ex Machina? Coercive Air Power in Bosnia and
Kosovo,” unpublished paper, School of Advanced Air Power Studies, Maxwell AFB,
Alabama, November 7, 1999, p. 10.
Chapter Five
ACCOMPLISHMENTS OF THE AIR WAR
87
88 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
RAND MR1365-5.1
Total: 23,315
PGM
(6,728)
29%
Non-PGM 71%
(16,587)
______________
5The qualification “U.S.-led” is appropriate here, considering that the Israeli Air Force
has made regular and highly effective use of UAVs over southern Lebanon for nearly
two decades, going back to the Beka’a valley air campaign of 1982.
Accomplishments of the Air War 89
______________
6The wing had most of the essential support assets on hand, so deploying squadrons
did not need to bring much by way of logistics overhead.
7Dale Eisman, “Over Balkans, It’s Beauty vs. the Beast,” Norfolk Virginian-Pilot, April
26, 1999.
8Vince Crawley, “B-2s See Combat over Yugoslavia,” Defense Week, March 29, 1999,
p. 6.
90 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Since only six of the nine available aircraft were actually used on
combat missions, the average turn time per aircraft was two days.9
There was never a shortage of capability to meet USEUCOM’s target-
ing needs, however. Some B-2s were turned in the time it took to
refuel them. The only reported case of a B-2 component having
failed during a combat mission was a malfunction of a rotary bomb
launcher, which was promptly repaired upon the aircraft’s return to
base.10 The chief maintenance drivers were said to have been the
aircraft’s low-observable treatment, its flight control system, its
synthetic-aperture radar, and engine accessory drives.
Each B-2 flew nonstop to its targets in its final Block 30 configuration
directly from Whiteman on 28- to 32-hour round-trip missions, de-
livering up to 16 global positioning system (GPS)-guided GBU-31
joint direct-attack munitions (JDAMs) from 40,000 ft, usually through
cloud cover, against enemy targets including hardened command
bunkers and air defense facilities. Those missions typically entailed
15-hour legs out and back, with two inflight refuelings per leg. Two
aircraft were launched on 15 nights and just a single aircraft on 19
nights. The aircrews quickly adjusted to these unprecedentedly long
missions and coped with them adequately. They also quickly
adapted to the demands of real-time targeting changes en route.
Although the USAF bomber community, by virtue of its traditional
nuclear focus, had long been predisposed to do things in a carefully
preplanned way, USAFE’s commander, General John Jumper, trav-
eled to Whiteman and personally talked to B-2 aircrews about the
need for rapid adaptability. After just a few hours of intense opera-
tor-to-operator brainstorming, any residual doubts some B-2 pilots
may have harbored regarding the merits of replacing traditional
cold-war practices with real-time improvisation as needed to meet
current demands were put to rest. The first time the ensuing air ef-
______________
9Of the nine available B-2s at Whiteman, one was kept aside for training, one was un-
dergoing final upgrades to Block 30 status, and one was in extensive maintenance.
“Missouri-to-Kosovo Flights for B-2 Not a Concern to Wing Commander,” Inside the
Air Force, July 2, 1999, p. 12.
10“B-2 Performed Better in Kosovo Than USAF Expected,” Inside the Pentagon, July 8,
1999, p. 11.
Accomplishments of the Air War 91
It bears emphasizing here that the B-2 did not merely drop weapons
preprogrammed to home in on assigned coordinates, but used its
onboard synthetic aperture radar (SAR) to take two successive
images of the target during its initial approach. By so doing, the B-2
was able to eliminate the largest target error source in the JDAM,
namely, the error in the exact location of the aim point in GPS space.
As a result, the B-2’s average miss distance with JDAM was less than
half the 13 meters stipulated for unassisted JDAMs.15
______________
11“Jumper on Air Power,” Air Force Magazine, July 2000, p. 43.
12Paul Richter, “B-2 Drops Its Bad PR in Air War,” Los Angeles Times, July 8, 1999.
13 Adam Hebert, “Air Force Follows Roadmap in Employment of Bombers Against
Serbia,” Inside the Air Force, April 2, 1999, p. 2.
14Barry D. Watts, “The EA-6B, E-8C, and B-2 in Operation Allied Force,” Northrop
Grumman Analysis Center briefing, Rosslyn, Virginia, May 8, 2000.
15Barry D. Watts, The Military Use of Space: A Diagnostic Assessment, Washington,
D.C., Center for Strategic and Budgetary Assessments, February 2001, p. 42.
92 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Since every B-2 mission, save one or two, benefited from dedicated
offboard electronic countermeasures (ECM) support and was flown
against less than top-of-the-line enemy defenses, it remains unclear
as to what extent the aircraft’s stealth properties were truly tested in
modern combat. However, by all accounts the aircraft was never
tracked by enemy radar, let alone shot at by enemy SAMs. Unlike all
other aircraft that flew combat missions in Allied Force, the B-2 op-
erated autonomously. It simply checked in with the ABCCC as it ap-
proached the target area, received a go/no-go code, and pressed
ahead to its assigned targets in radio silence. If a target change was
required en route, the Combined Air Operations Center (CAOC)
could pass essential information to ingressing B-2 aircrews as much
as an hour and 45 minutes before the aircraft’s scheduled time on
target (TOT). That ability to select new targets while airborne en-
abled the aircraft to take out some enemy SA-3s and their radars
shortly after they were located and identified by allied sensors.18 The
______________
16Watts,” The EA-6B, E8C, and B-2 in Operation Allied Force.”
17Colonel Tony Imondi, 509th Operations Group commander, quoted in Bill Sweet-
man, “B-2 Is Maturing into a Fine Spirit,” Jane’s International Defense Review, May
2000.
18Brigadier General Randy Gelwix, USAF, “Oral Histories Accomplished in Conjunc-
tion with Operation Allied Force/Noble Anvil.”
Accomplishments of the Air War 93
As the air war unfolded, former Secretary of the Air Force Donald
Rice observed that the B-2, although one of the most controversial
weapons in the U.S. inventory, was “proving to be the nation’s single
most cost-effective attack aircraft.” 19 Rice further pointed out that
the much-derided stealth treatments on the aircraft had proven
themselves durable and reliable and that the aircraft had been con-
sistently flying through inclement weather and returning home in
serviceable condition. As for identified shortcomings, the B-2 was
found to need a direct satellite link to national intelligence agencies
to provide its crew with a more current picture of the electronic bat-
tlefield so that the aircraft could be rerouted in near-real time to
avoid any pop-up SAM threats that might have been detected after it
had taken off. It also became apparent, at least to some observers,
that the 509th Bomb Wing’s crew ratio of two two-pilot crews per air-
craft might need to be increased to four crews, or else that provisions
might need to be made for future combat contingencies to allow the
B-2 to operate out of airfields closer to the battlespace in the interest
of reducing mission times.20
______________
19Donald B. Rice, “No Stealth to Pentagon’s Bias Against the B-2,” Los Angeles Times,
May 9, 1999.
20David A. Fulghum, “Lessons Learned May Be Flawed,” Aviation Week and Space
Technology, June 14, 1999, p. 205. A serious limiting factor affecting the first of these
suggested solutions is that doubling the B-2’s crew ratio would require either doubling
the number of training sorties and hours flown by the Air Force’s limited B-2 inventory
or reducing the number of sorties and flying hours made available for each B-2 crew
member—to a point where their operational proficiency and expertise would be unac-
ceptably compromised. Alternatively, the Air Force is now taking a close look at using
RAF Fairford, England, and the island bases of Diego Garcia and Guam as forward
staging areas from which to conduct B-2 operations in future regional contingencies
worldwide.
94 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
UAV EMPLOYMENT
Also for the first time in American combat experience, UAVs offered
commanders and planners the frequent advantage of real-time video
imagery without any accompanying danger of aircrew losses. Some
UAVs were flown as low as 1,000 ft above VJ troop positions to gather
real-time imagery, which, in turn, occasionally enabled prompt and
effective attacks by A-10s and F-16s against the often fleeting targets.
Several UAVs were lost when commanders requested closer looks,
forcing the drones to descend into the lethal envelopes of Serb AAA
and man-portable air defense systems (MANPADS). These losses did
not evoke great concern, however, since the UAVs were intentionally
sent out on missions that were known ahead of time to be especially
risky, including highly classified missions to collect and downlink
evidence on Serb atrocities.22
______________
21William M. Arkin, “In Praise of Heavy Bombers,” Bulletin of the Atomic Scientists,
July–August 1999. Another 218 U.S. and British TLAMs were fired during Operation
Allied Force.
22“Despite Losses, Backers Say Unmanned Systems Excelling Over Kosovo,” Inside the
Pentagon, June 10, 1999, p. 1.
Accomplishments of the Air War 95
sions, such as those from cell phones and portable radios operated
by enemy ground troops. 23
______________
23John D. Morrocco, David Fulghum, and Robert Wall, “Weather, Weapons Dearth
Slow NATO Strikes,” Aviation Week and Space Technology, April 5, 1999, p. 29.
24“Air Force Reluctant to Deploy All-Weather Predator UAVs to Balkans,” Inside the
Air Force, April 2, 1999, p. 1. Another concern had to do with a larger requirements
debate within the Air Force over whether UAVs developed under a fast-track acquisi-
tion process, as was Predator, should be managed like a more expensive fighter pro-
gram.
25Jane Perlez, “Serbs Try to Empty Disputed Province, NATO Aides Assert,” New York
Times, March 29, 1999.
96 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
The U.S. Army’s Hunter UAVs operated from the Skopje airfield in
Macedonia, with their first operational mission into Kosovo taking
place on April 4. Hunter imagery was first downlinked to ground
controllers in Skopje and then forwarded either to the CAOC in Vi-
cenza, Italy, or to NATO headquarters in Belgium and to the Pen-
tagon as appropriate.28 Normally used as a corps asset, Hunter in
this instance transmitted real-time video imagery via orbiting satel-
lites and downlinked it directly to the Joint Broadcast System in the
United States, which then transmitted it to the CAOC, making for
only a one-second delay. Its targets were normally objects of tactical
interest against which commanders would not risk a manned air-
craft, such as artillery emplacements and dispersed VJ units in the
KEZ, which had organic self-protection air defense assets. Much like
Predator, Hunter flew whenever the weather allowed. It often would
loiter in the vicinity of hot targets to observe munitions impacts and
provide real-time BDA.29
Both Predator and Hunter operators soon discovered that better sen-
sors were needed for the drones to identify ground targets positively
from above 8,000 ft. They also learned that better integration of UAV
and manned aircraft operations was essential for minimizing the
______________
26“Jumper on Air Power,” p. 42.
27One problem pointed up by this mode of operation was the slow flying speed of the
aircraft. At a maximum airspeed of only 70 nautical miles per hour, Predator typically
required considerable time to get to a previously located target candidate, by which
time the latter may have moved to a new location.
28Elizabeth Becker, “They’re Unmanned, They Fly Low, and They Get the Picture,”
New York Times, June 3, 1999.
29 Tim Ripley, “Task Force Hunter,” World Air Power Journal, Winter 1999/2000,
p. 122.
Accomplishments of the Air War 97
After Allied Force ended, General Jumper revealed that had combat
operations continued into the summer, the USAF would have started
employing a new tactic whereby Predators equipped with laser des-
ignators would have been flown under the weather near enemy tar-
gets to designate those targets for LGBs once the latter had been re-
leased by allied fighters flying at safer altitudes above the cloud
cover. Jumper further disclosed that UAVs, having successfully un-
dergone a rigorous operational shakedown over Kosovo, would in the
future be used more in the targeting loop than in the intelligence
collection loop—patrolling aggressively and making the most of their
extended loiter time to seek out and identify hidden targets.32
______________
30David A. Fulghum, “Joint STARS May Profit from Yugoslav Ops,” Aviation Week and
Space Technology, July 26, 1999, p. 74.
31William M. Arkin, “Top Air Force Leaders to Get Briefed on Serbia Air War Report,”
Defense Daily, June 13, 2000, p. 1. For further details on UAV operations, see Lieu-
tenant Commander J. D. Dixon, “UAV Employment in Kosovo: Lessons for the Opera-
tional Commander,” paper submitted to the Naval War College, Newport, Rhode
Island, February 8, 2000.
32David A. Fulghum, “Kosovo Conflict Spurred New Airborne Technology Use,” Avia-
tion Week and Space Technology, August 23, 1999, p. 30.
98 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
At least five notable space success stories came out of the Allied
Force experience. The first was the effective use of the Multisource
Tactical System (MSTS) on the B-52 and B-1, which gave bomber
crews real-time situation awareness updates. The system had existed
before but had never previously been used in combat. The second
major success story was the highly successful use of GPS-guided
munitions described earlier, most notably JDAM on the B-2 and the
Navy’s TLAM II. Third was the use of the Defense Support Program
(DSP) satellite constellation for providing real-time battle damage
indications (BDI) as an input into the BDA process. New procedures
toward that end were created and refined for Allied Force that had
never before been used.35 Fourth, the Hook 112 survival radio was
available for use by U.S. aircrews, making an important new role for
space-enabling systems in CSAR. 36 Finally, command and control
______________
33Craig Covault, “Military Space Dominates Air Strikes,” Aviation Week and Space
Technology, March 29, 1999, pp. 31–32.
34Peter Grier, “The Investment in Space,” Air Force Magazine, February 2000, p. 50.
35On the other hand, cockpit multifunction display videotapes showing successfully
impacting munitions were not used in the BDA process by the Joint Analysis Center at
RAF Molesworth, resulting in numerous revisits to targets that were already known by
attacking pilots to have been struck before to good effect. Conversation with Lieu-
tenant Colonel Ray Dissinger, Aviano AB, Italy, June 12, 2000.
36 The Hook 112 was developed by the Air Force for use between downed aircrew
members and CSAR forces to eliminate a problem presented by the previous survival
radio, which allowed enemy monitors to locate the downed crewmember’s position by
Accomplishments of the Air War 99
_____________________________________________________________
triangulating on the relatively lengthy voice exchanges required to coordinate a rescue
by CSAR teams. The Hook 112 communicates the downed crewmember’s position by
means of an encrypted burst transmission that denies enemy monitors the ability to
triangulate. A GPS receiver incorporated in the Hook 112 automatically transmits the
crewmember’s exact location, along with any coded transmissions the downed
crewmember may wish to communicate. Major General Gary Dylewski, “The USAF
Space Warfare Center: Bringing Space to the Warfighter,” in Peter L. Hays et al., eds,
Spacepower for a New Millennium: Space and U.S. National Security, New York,
McGraw-Hill, 2000, p. 96.
37“Space Support to Operation Allied Force: Preliminary Lessons Learned,” briefing
to the author by Colonel Robert Bivins, director of operations, U.S. Air Force Space
Warfare Center, Schriever AFB, Colorado, February 25, 2000.
38Major General Robert Hinson, commander, 14th Air Force, “Space Doctrine Lessons
from Operation Allied Force,” command briefing, Vandenberg AFB, California,
December 16, 1999.
This page intentionally blank
Chapter Six
FRICTION AND OPERATIONAL PROBLEMS
Although NATO’s use of air power in Allied Force must, in the end, be
adjudged a success, some troubling questions arose well before the
air war’s favorable outcome over a number of unexpected and dis-
concerting problems encountered along the way. Some of those
problems, most notably in the area of what air planners came to call
“flex” targeting of elusive VJ troops on the move in Kosovo, were ar-
guably as much a predictable result of prior strategy choices as a re-
flection of any inherent deficiencies in the air weapon itself.1 Of
more serious concern were identified shortcomings that indicated
needed fixes in the realm of tactics, techniques and procedures, and,
in some cases, equipment. Beyond the problem of locating, identify-
ing, and engaging dispersed and hidden light infantry targets, the
shortcomings arousing the greatest consternation included assessed
deficiencies in SEAD, excessively lengthy information and intelli-
gence cycle time, inadvertent civilian casualties, and some serious
deficiencies in alliance interoperability. Also of special concern were
the many problems spotlighted by the U.S. Army’s plagued deploy-
ment of its AH-64 Apache helicopters to Albania and the full extent of
U.S. global military overcommitment that the Allied Force experi-
ence brought to light.
______________
1The “flex” targeting effort entailed the launching of combat aircraft without specific
assigned target locations and coordinates, although tasked to seek out various classes
of targets, either through free search or upon being directed to a specific area of
known or suspected enemy activity by the CAOC or an airborne forward air controller.
101
102 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
The Serbs, however, kept their SAMs defensively dispersed and op-
erating in an emission control (EMCON) mode, prompting concern
that they were attempting to draw NATO aircraft down to lower alti-
tudes where they could be more easily engaged. Before the initial
strikes, there were reports of a large-scale dispersal of SA-3 and SA-6
batteries from nearly all of the regular known garrisons. The under-
standable reluctance of enemy SAM operators to emit and thus ren-
der themselves cooperative targets made them much harder to find
and attack, forcing allied aircrews to remain constantly alert to the
radar-guided SAM threat throughout the air war. 2 It further had the
effect of denying some high-risk targets for a time, increasing force
package size, and increasing overall SEAD sortie requirements.
______________
2Dana Priest, “NATO Unlikely to Alter Strategy,” Washington Post, March 26, 1999.
3Dana Priest, “NATO Pilots Set to Confront Potent Foe,” Washington Post, March 24,
1999.
Friction and Operational Problems 103
which meant that the Serbs knew when attacks were coming most of
the time.4 In other cases, the cumbersome command and control ar-
rangements and the need for prior CAOC approval before fleeting
pop-up IADS targets detected by Rivet Joint or other allied sensors
could be attacked resulted in many lost opportunities and few hard
kills of enemy SAM sites.
______________
4General John Jumper, USAF, “Oral Histories Accomplished in Conjunction with Op-
eration Allied Force/Noble Anvil.”
5 Robert Wall, “Sustained Carrier Raids Demonstrate New Strike Tactics,” Aviation
Week and Space Technology, May 10, 1999, p. 37.
104 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
6Robert Wall, “Airspace Control Challenges Allies,” Aviation Week and Space Technol-
ogy, April 26, 1999, p. 30.
7Tim Ripley, “Viper Weasels,” World Air Power Journal, Winter 1999/2000, p. 102. The
standard F-16CJ weapons loadout was two AGM-88 HARMs and four AIM-120 ad-
vanced medium-range air-to-air missiles (AMRAAMs).
8Richard J. Newman, “In the Skies over Serbia,” U.S. News and World Report, May 24,
1999, p. 24. It bears noting here that 10 or more pilots operating in a target area might
report an observed SAM shot as ballistic, while the one pilot on whose helmet the
Friction and Operational Problems 105
Indeed, the SAM threat to NATO’s aircrews was far more pronounced
and harrowing than media coverage typically depicted, and aggres-
sive jinking and countermaneuvering against airborne SAMs was fre-
quently necessary whenever the Serbs sought to engage NATO air-
craft. The Supreme Allied Commander in Europe, U.S. Army General
Wesley Clark, later reported that there had been numerous instances
of near-misses involving enemy SAM launches against NATO aircraft,
and General Jumper added that a simple look at cockpit display
videotapes would show that “those duels were not trivial.”9 From the
very start of NATO’s air attacks, Serb air defenders also sought to
sucker NATO aircrews down to lower altitudes so they could be
brought within the lethal envelopes of widely proliferated MANPADS
and AAA systems. A common Serb tactic was to fire on the last air-
craft in a departing strike formation, perhaps on the presumption
that those aircraft would be unprotected by other fighters, flown by
less experienced pilots, and low on fuel, with a consequent limited
latitude to countermaneuver.
_____________________________________________________________
missile was figuratively guiding would be actively reacting to it. Shortly thereafter, 10
pilots would recover to widely dispersed home bases and report nonthreatening bal-
listic launches, while only one would return with the evidence of a guided shot. This
drove a perception among Allied Force leaders that “most” of the SAM shots observed
were ballistic. Once all the pertinent information was fused and duplicate reporting
was factored out, however, it turned out that a substantial number of SAM launches
(perhaps as many as a third) were guided. Comments on an earlier draft by Hq
USAFE/IN, May 18, 2001.
9Cited in “Ground Troops Lauded,” European Stars and Stripes, August 6, 1999, and
“Jumper on Air Power,” Air Force Magazine, July 2000, p. 41.
106 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
In all, well over half of the HARM shots taken by allied SEAD aircrews
were preemptive targeting, or so-called PET, shots, with a substantial
number of these occurring in the immediate Belgrade area.12 Many
HARM shots, however, were reactive rather than preplanned, made
in response to transitory radar emissions as they were detected.13
______________
10 Further mitigating this constraint, the limited surveillance range of Joint STARS
caused by interposed ridge lines restricted E-8 operations primarily with regard to
Kosovo, which harbored only a limited SAM threat (only one of the 5 SA-6 regiments
and no SA-2s or SA-3s). Most of the enemy IADS targets were assessed to lie outside
Kosovo. Moreover, the U-2 and Rivet Joint typically performed well and did not suffer
the same problems that sometimes plagued the E-8. Comments on an earlier draft by
Hq USAFE/IN, May 18, 2001.
11Wall, “Airspace Control Challenges Allies.”
12Brigadier General Randy Gelwix, USAF, “Oral Histories Accomplished in Conjunc-
tion with Operation Allied Force/Noble Anvil.”
13Wall, “Airspace Control Challenges Allies,” p. 30.
Friction and Operational Problems 107
transportation for mobility and towed AAA tended to bog down when
driven off prepared surfaces and into open terrain. NATO pilots
therefore studiously avoided flying down roads and crossed them
when necessary at 90-degree angles to minimize their exposure time.
By remaining at least 5 km from the nearest road, they often were
able to negate the AAA threat, albeit at the cost of making it harder to
spot moving military vehicles.
One problem with such DEAD attempts was that the data cycle time
had to be short enough for the attackers to catch the emitting radars
before they moved on to new locations. One informed report ob-
served that supporting F-16CJs were relatively ineffective in the re-
active SEAD mode because the time required for them to detect an
impending launch and get a timely HARM shot off to protect a striker
______________
14The AGM-130 could be fired from a standoff range of up to 30 nautical miles. It fea-
tured GPS guidance, enhanced by terminal homing via man in the loop through live
video feed data-linked to the attacking aircraft from the guiding weapon.
15The Block 50/52 F-16CJs used for defense suppression were equipped to carry the
AGM-65 Maverick missile, but they did not employ that munition in Allied Force be-
cause the pilots, given their predominant focus on making the most of the AGM-88
HARM, had not sufficiently trained for its use.
16Gelwix, “Oral Histories.” JSOW was employed only infrequently during Allied Force.
Many of the targets assigned to the Navy were inappropriate for attack by the AGM-
154’s cluster-bomb variant because of collateral damage concerns and the lengthy
timelines associated with attacks against mobile targets and with the munition’s lack
of a precise impact timeline. William M. Arkin, “Fleet Praises JSOW, Lists Potential
Improvements,” Defense Daily, April 26, 2000.
108 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
invariably exceeded the flyout time of the SAM aimed at the targeted
aircraft. As a result, whenever attacking fighters found themselves
engaged by a SAM, they were pretty much on their own in defeating
it. That suggested to at least some participating aircrews the value of
having a few HARMs uploaded on selected aircraft in every strike
package so that strikers could protect themselves as necessary with-
out having to depend in every case on F-16CJ or EA-6B support.17
In all, more than 800 SAMs were reported to have been fired at NATO
aircraft, both manned and unmanned, over the course of the 78-day
air war, including 477 SA-6s and 124 confirmed man-portable in-
frared missiles (see Figure 6.1 for a depiction of reported enemy SAM
launches by type).19 A majority of the fixed SAMs were fired without
any radar guidance. Yet despite that expenditure of assets, only two
NATO aircraft, an F-117 and an F-16, were shot down by enemy fire,
although another F-117 sustained light damage from a nearby SA-3
detonation and two A-10s were hit by enemy AAA fire but not
downed. 20 There also were two reported cases of short-range
infrared (IR)-guided missiles hitting A-10s, one of which apparently
struck the bottom of the aircraft, defused itself, and bounced off
______________
17Lieutenant Colonel Philip C. Tissue, USMC, “21 Minutes to Belgrade,” Proceedings,
U.S. Naval Institute, September 1999, p. 40.
18Michael R. Gordon, “NATO to Hit Serbs from 2 More Sides,” New York Times, May
11, 1999.
19“AWOS Fact Sheet,” Hq USAFE/SA, December 17, 1999. See also William M. Arkin,
“Top Air Force Leaders to Get Briefed on Serbia Air War Report,” Defense Daily, June
13, 2000, p. 1.
20David A. Fulghum, “Kosovo Report to Boost New JSF Jamming Role,” Aviation Week
and Space Technology, August 30, 1999, p. 22.
Friction and Operational Problems 109
RAND MR1365-6.1
Total: 815
Unknown
(26)
IR/MANPAD
(124) SA-3
(188)
SA-6
(477)
harmlessly. 21 At least 743 HARMs were fired by U.S. and NATO air-
craft against the radars supporting these enemy SAMs (Figure 6.2
provides a detailed breakout of HARM expenditure by target type).22
Yet enough of the Serb IADS remained intact to require NATO
fighters to operate above the 15,000-ft hard deck for most of the air
effort. The main reason for this requirement was the persistent AAA
and MANPADS threat. Although the older SA-7 could be effectively
______________
21“Washington Outlook,” Aviation Week and Space Technology, September 20, 1999,
p. 25.
22“AWOS Fact Sheet,” Hq USAFE/SA, December 17, 1999.
110 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
RAND MR1365-6.2
Total: 743
Unknown
(20)
Early warning
radars
(125)
SA-2
(1)
SA-6
(389)
SA-3
(208)
In the end, as noted above, only two aircraft (both American) were
brought down by enemy SAM fire, thanks to allied reliance on elec-
tronic jamming, the use of towed decoys, and countertactics to
negate enemy surface-to-air defenses.23 However, NATO never fully
succeeded in neutralizing the Serb IADS, and NATO aircraft operat-
ing over Serbia and Kosovo were always within the engagement en-
velopes of enemy SA-3 and SA-6 missiles—envelopes that extended
______________
23 In all, 1,479 ALE-50 towed decoys were expended by U.S. aircraft during Allied
Force.
Friction and Operational Problems 111
______________
24Comments on an earlier draft by Hq USAFE/IN, May 18, 2001.
25Tim Ripley, “‘Serbs Running Out of SAMs,’ Says USA,” Jane’s Defense Weekly, June 2,
1999.
26Interview with Lieutenant General Michael Short, USAF, PBS Frontline, “War in Eu-
rope,” February 22, 2000. Serb IADS operators may have been able to trade short-term
effectiveness for longer-term survivability because allied aircraft were typically unable
to find and successfully attack VJ fielded forces and other mobile ground targets. Had
they been able to do so and to kill VJ troops in large numbers, the VJ’s leadership
would have insisted on a more aggressive air defense effort. That would have enabled
NATO to kill more SAMs, but at the probable cost of more friendly aircraft lost. I am
indebted to my RAND colleague John Stillion for this insight.
112 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
The dearth of enemy radar-guided SAM activity may also have been
explainable, at least in part, by reports that the Air Force’s Air Com-
bat Command had been conducting information operations by in-
serting viruses and deceptive communications into the enemy’s
computer system and microwave net.27 Although it is unlikely that
U.S. information operators were able to insert malicious code into
enemy SAM radars themselves, General Jumper later confirmed that
Operation Allied Force had seen the first use of offensive computer
warfare as a precision weapon in connection with broader U.S. in-
formation operations against enemy defenses. As he put it, “we did
more information warfare in this conflict than we have ever done
before, and we proved the potential of it.” Jumper added that al-
though information operations remained a highly classified and
compartmented subject about which little could be said, the Kosovo
experience suggested that “instead of sitting and talking about great
big large pods that bash electrons, we should be talking about mi-
crochips that manipulate electrons and get into the heart and soul of
systems like the SA-10 or the SA-12 and tell it that it is a refrigerator
and not a radar.”28 Such pioneering attempts at offensive cyber-
warfare pointed toward the feasibility of taking down SAM and other
defense systems in ways that would not require putting a strike pack-
age or a HARM missile on critical nodes to neutralize them.
______________
27David A. Fulghum, “Serb Threat Subsides, but U.S. Still Worries,” Aviation Week and
Space Technology, April 12, 1999, p. 24.
28“Jumper on Air Power,” p. 43.
Friction and Operational Problems 113
Fortunately for NATO, the Serb IADS did not include the latest-
generation SAM equipment currently available on the international
arms market. There were early unsubstantiated reports, repeatedly
denied by the Russian Ministry of Foreign Affairs, that several weeks
before the start of the bombing effort, Russia had provided
Yugoslavia with elements of between six and ten S-300PM (NATO
code-name SA-10) long-range SAM systems, which had been deliv-
ered without their 36D6 Clam Shell low-altitude acquisition radars.30
Had those reports been valid, even the suspected presence of SA-10
and SA-12 SAMs in the enemy IADS inventory would have made life
far more challenging for attacking NATO aircrews. Milosevic
reportedly pressed the Russians hard for such equipment repeatedly,
without success. Deputy Secretary of State Strobe Talbott later
stated that the Yeltsin government had been put on the firmest
notice by the Clinton administration that any provision of such
cutting-edge defensive equipment to Yugoslavia would have had a
“devastating” effect on Russian-American relations.31
All of this raised basic questions about the adequacy of U.S. SEAD
tactics and suggested a need for better real-time intelligence on
mobile enemy IADS assets and a means of getting that information
to pilots quickly enough for them to act on it, as well as for greater
standoff attack capability.32 The downings of both the F-117 and F-
16 were attributed to breakdowns in procedures aimed at detecting
______________
29David A. Fulghum, “Yugoslavia Successfully Attacked by Computers,” Aviation Week
and Space Technology, August 23, 1999, pp. 31–34.
30Zoran Kusovac, “Russian S-300 SAMs ‘In Serbia,’” Jane’s Defense Weekly, August 4,
1999.
31Michael Ignatieff, Virtual War: Kosovo and Beyond, New York, Henry Holt and
Company, Inc., 2000, p. 109.
32For example, the SA-10 and SA-12, now available on the international arms market
for foreign military sale, are lethal out to a slant range of some 80 nautical miles, five
times the killing reach of the earlier-generation SA-3 (David A. Fulghum, “Report Tal-
lies Damage, Lists U.S. Weaknesses,” Aviation Week and Space Technology, February
14, 2000, p. 34). One SA-10/12 site in Belgrade and one in Pristina could have provided
defensive coverage of all of Serbia and Kosovo, as well as threatened Compass Call and
the ABCCC operating outside enemy airspace.
114 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
One of the first signs of this insidious trend cropped up as far back as
August 1990, when half of the Air Force’s ECM pods being readied for
deployment to the Arabian peninsula for Desert Storm were found to
have been in need of calibration or repair. Among numerous later
sins of neglect with respect to electronic warfare (EW) were Air Force
decisions to make operational readiness inspections (ORIs) and
Green Flag EW training exercises less demanding, decisions that nat-
urally resulted in an atrophying of the readiness inspection and re-
porting of EW units, along with a steady erosion of EW experience at
the squadron level. “Now,” said the Air Force general cited above,
“they only practice reprogramming [of radar warning receivers] at
the national level. Intelligence goes to the scientists and says the sig-
nal has changed. Then the scientists figure out the change for the
[ECM] pod and that’s it. Nobody ever burns a new bite down at the
wing.”34
During the years since Desert Storm, the response time for SEAD
challenges has become longer, not shorter, owing to an absence of
adequate planning and to the disappearance of a talent pool of Air
Force leaders skilled in EW. One senior Air Force Gulf War veteran
complained that “we used to have an XOE [operational electronic
______________
33David A. Fulghum, “NATO Unprepared for Electronic Combat,” Aviation Week and
Space Technology, May 10, 1999, p. 35. A thorough and detailed account of the many
problems and concerns identified and highlighted with respect to the USAF’s current
SEAD and electronic warfare repertoire is contained in the summary report of an Air
Force–commissioned study by RAND’s Natalie Crawford and seven senior retired Air
Force general-officer electronic warfare experts, “USAF EW Management Process
Study,” briefing charts, October 1, 1999.
34Fulghum, “NATO Unprepared for Electronic Combat,” p. 35.
Friction and Operational Problems 115
warfare] branch in the Air Staff. That doesn’t exist any more. We
used to reprogram [ECM] pods within the wings. They don’t really
do that any more.” During a subsequent colloquium on the air war
and its implications, former Air Force chief of staff General Michael
Dugan attributed these problems to the Air Force’s having dropped
the ball badly in 1990, when it failed to “replace a couple of senior
officers in the acquisition and operations community who [oversaw]
the contribution of electronic combat to warfighting output. The
natural consequence was for this resource to go away.”35
______________
35“Washington Outlook,” Aviation Week and Space Technology, August 23, 1999, p. 27.
36Robert Wall, “SEAD Concerns Raised in Kosovo,” Aviation Week and Space Technol-
ogy, June 26, 1999, p. 75.
37“Washington Outlook,” Aviation Week and Space Technology, June 7, 1999, p. 23.
116 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
side result of this neglect of the EW mission by the Air Force was that
maintenance technicians could no longer reprogram quickly (that is,
in 24 hours or less) ECM pods and radar warning receivers to counter
newly detected enemy threats. That problem first arose in 1998,
when several planned U-2 penetrations into hostile airspace had to
be canceled at the last minute because USAF radar warning systems
could not recognize some IADS signals emanating from Iraq and
Bosnia.
______________
38“Washington Outlook,” Aviation Week and Space Technology, May 24, 1999, p. 27.
Friction and Operational Problems 117
combat search and rescue team using MH-53 Pave Low and MH-60
Pave Hawk helicopters, and directed by a flight of A-10s.39
______________
39Although some criticism was voiced afterward as to how CSAR had been shown to
be “broken” because of problems that cropped up during the rescue operation
(apparently, one of the helicopters was forced to disengage, refuel, and penetrate en-
emy airspace a second time before it could find and finally retrieve the downed pilot),
genuine acts of heroism were displayed during the mission. It ended up a brilliant
success and had the welcome effect of turning a propaganda coup for Milosevic al-
most instantly into a propaganda coup for NATO. On the criticism expressed, see
Rowan Scarborough, “Air Force Search and Rescue Operations Called ‘Broken,’”
Washington Times, September 13, 1999.
40James Peltz and Jeff Leeds, “Stealth Fighter’s Crash Reveals a Design’s Limits,” Los
Angeles Times, March 30, 1999.
41“Washington Outlook,” Aviation Week and Space Technology, May 3, 1999, p. 21.
Asked whether the aircraft’s loss was caused by a failure to observe proper lessons
from earlier experience, Hawley added: “That’s an operational issue that is very
warm.”
118 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Last was the reported fact that F-117s operating out of Aviano had
previously flown along more or less the same transit routes for four
nights in a row because of a SACEUR ban on overflight of Bosnia to
______________
42Eric Schmitt, “Shrewd Serb Tactics Downed Stealth Jet, U.S. Inquiry Shows,” New
York Times, April 11, 1999. In subsequent testimony before the Senate Armed Services
Committee, Secretary of the Air Force F. Whitten Peters did confirm that the aircraft
had been downed by enemy SAMs. See Vince Crawley, “Air Force Secretary Advocates
C-130, Predators,” Defense Week, July 26, 1999, p. 2.
43See David A. Fulghum and William B. Scott, “Pentagon Gets Lock on F-117 Shoot-
down,” Aviation Week and Space Technology, April 19, 1999, pp. 28–30, and Paul
Beaver, “Mystery Still Shrouds Downing of F-117A Fighter,” Jane’s Defense Weekly,
September 1, 1999.
Friction and Operational Problems 119
avoid jeopardizing the Dayton accords. That would have made their
approach pattern into Yugoslav airspace predictable. Knowing from
which direction the F-117s would be coming, Serb air defenders
could have employed low-frequency radars for the best chance of
getting a snap look at the aircraft. Former F-117 pilots and several
industry experts acknowledged that the aircraft is detectable by such
radars when viewed from the side or from directly below. U.S. offi-
cials also suggested that the Serbs may have been able to get brief
nightly radar hits while the aircraft’s weapons bay doors were fleet-
ingly open.
______________
44To bolster their case, some noted that when an F-117 had crashed earlier at an air
show near Baltimore in 1998, the Air Force had thoroughly sanitized the area and
hauled off the wreckage to prevent its most sensitive features from being compro-
mised.
45Vago Muradian, “Stealth Compromised by Not Destroying F-117 Wreckage,” De-
fense Daily, April 2, 1999.
46Ibid.
120 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
through with the attempt because they could not have located it
quickly enough to attack it before it was surrounded by civilians and
the media. 47 Those issues aside, whatever the precise explanation
for the downing, it meant not merely the loss of a key U.S. combat
aircraft but the dimming of the F-117’s former aura of invincibility,
which for years had been of incalculable psychological value to the
United States.
______________
47On April 2, the Yugoslav government announced its intention to hand over pieces of
the downed F-117 to Russian authorities. Robert Hewson, “Operation Allied Force:
The First 30 Days,” World Air Power Journal, Fall 1999, p. 18. For the record, it should
be noted that USAF F-15Es were immediately put on alert to destroy the wreckage
with AGM-130s after the F-117 downing was confirmed, but by the time the wreckage
location could be positively determined, CNN was on the scene and collateral damage
issues precluded the attack. Comments on an earlier draft by Hq USAF/XOXS, July 9,
2001.
Friction and Operational Problems 121
Indeed, the opportunity to get at fielded enemy ground units with air
power alone had been essentially lost by NATO even before Opera-
tion Allied Force commenced. As General Jumper later recalled,
during the Rambouillet talks in early March 1999, “we watched
40,000 Serbian troops mass north of Kosovo, we watched them infil-
trate down into Kosovo, we watched heavy armor come down into
there, all under the umbrella of the peace conference, and we
weren’t able to react.” 48 Once those forces had completed their
massing on March 15 and had begun a substantial incursion into
Kosovo, any chance for allied air power to be significantly effective
against them promptly disappeared. Once safely dispersed, VJ units
simply turned off the engines of their tanks and other vehicles to save
fuel, hid their vehicles in barns, churches, forests, and populated ar-
eas, hunkered down, and hoped to wait the air effort out. By the end
of April, General Clark frankly conceded that after six weeks of
bombing, there were more VJ, MUP, and Serb paramilitary forces in
Kosovo than there had been when Allied Force began. That attested
powerfully to the latter’s near-total ineffectiveness, at least up to that
point, in halting the Serbian ethnic cleansing rampage throughout
Kosovo.
______________
48General John Jumper, USAF, “Oral Histories Accomplished in Conjunction with Op-
eration Allied Force/Noble Anvil.”
49Robert Wall, “Joint STARS Changes Operational Scheme,” Aviation Week and Space
Technology, May 3, 1999, pp. 25–27.
122 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
50I am indebted to my colleague John Stillion for developing these points.
124 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
51Personal communication to the author from Price Bingham, Northrop Grumman
Corporation, Melbourne, Florida, December 20, 1999.
Friction and Operational Problems 125
Even on clear days, another factor preventing the kill box system
from being as effective as it might otherwise have been was the tight
rules-of-engagement regime that had been imposed after the
Djakovica incident (see below), in which more than 60 ethnic Alba-
nian refugees were reportedly killed in an attack by USAF F-16s
against what was thought to have been a VJ troop convoy. These re-
strictions had a far greater inhibiting influence on the effectiveness of
NATO’s flexible targeting efforts than the oft-cited 15,000-ft altitude
floor which NATO’s aircrews had been directed to observe. Unless
an object of interest was clearly determined to be a valid military tar-
get, such as a VJ tank operating in the open, pilots had to get clear-
ance for any attack from the CAOC, with General Short himself often
______________
52Lieutenant Colonel L. T. Wight, USAF, “What a Tangled Web We Wove: An After-
Action Assessment of Operation Allied Force’s Command and Control Structure and
Processes,” unpublished paper, no date, p. 12. Colonel Wight was a member of the
C-5 Strategy Cell at the CAOC.
53Ibid.
126 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Last, and perhaps as decisive as any single other factor, VJ forces ag-
gressively avoided making themselves easy targets for NATO air at-
tacks. Indeed, digging in and hunkering down for defensive attrition
warfare had lain at the heart of Yugoslav operational doctrine ever
since the days of partisan operations against the Wehrmacht in
World War II. Whenever General Clark would say, “You’ve got to get
them in their assembly areas,” the reply typically was: “These guys
aren’t assembling!”55 RAF Harrier GR. Mk 7 pilots operating in kill
boxes over Kosovo reported that “there was nothing moving around
at all during the daytime,” adding that when Clark “got up and said
knocking out five tanks was a good day for NATO, he [was] telling it
straight. On some days we couldn’t find any tanks.”56 Even with the
aid of binoculars, the ground below often seemed devoid of life to
NATO aircrews orbiting overhead at 15,000 ft. This was the pre-
dictable result of trying to engage an enemy who had no need to
shoot, move, or expose his position, thanks to the absence of a cred-
ible NATO ground threat.
To be sure, there were some notable bright spots in NATO’s air effort
against VJ forces in Kosovo. To cite one example, in those rare in-
stances in which enemy armor and other targets exposed themselves
to attack from the air, the upgraded AGM-65G2 Maverick air-to-
ground missile generally performed very effectively. The effective-
ness rate for older Mavericks was lower, but still reportedly higher
than 90 percent. 57 Also, both U-2 imagery and pictures provided by
the Navy’s F-14 equipped with TARPS (Tactical Air Reconnaissance
______________
54Tim Ripley, “Harriers over the Kosovo ‘Kill Boxes,’” World Air Power Journal, Winter
1999/2000, p. 100.
55Brigadier General Randy Gelwix, USAF, “Oral Histories Accomplished in Conjunc-
tion with Operation Allied Force/Noble Anvil.”
56Ripley, “Harriers over the Kosovo ‘Kill Boxes.’”
57Robert Wall, “Maverick Fix Tested in Kosovo,” Aviation Week and Space Technology,
September 6, 1999, pp. 88–89.
Friction and Operational Problems 127
Pod System) later proved useful to the CAOC in what the Cohen-
Shelton after-action report to Congress called “several” instances in-
volving the rapid retargeting of NATO aircraft to new targets.58
______________
58 Secretary of Defense William S. Cohen and Chairman of the Joint Chiefs of Staff
General Henry H. Shelton, Kosovo/Operation Allied Force After-Action Report, Wash-
ington, D.C., Department of Defense, Report to Congress, January 31, 2000, p. 58.
59The airfield itself offered an 8,200-ft runway and a tactical air navigation (TACAN)
system enabling the aircraft to fly instrument approaches, but it lacked a ready com-
munications link to the CAOC in Vicenza and also needed more fuel trucks, as well as
runway arresting gear in the event of wet runways and aircraft emergencies. The latter
were shipped in and quickly became a welcome presence because high-gross-weight
landings in heavy rain proved to be routine.
60As one downside aspect worth noting in this respect, numerous aircrews later indi-
cated that night-vision goggles often provided them with too much information be-
cause they were capable of picking up infrared events as far as 100 miles away.
61 For further details, see Margaret Bone, “Kodak Moments in Kosovo,” The Hook,
Spring 2000, pp. 29–31.
128 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
62Tissue, “21 Minutes to Belgrade,” pp. 38–40.
63Conversation with Major General P. J. M. Godderij, deputy commander in chief,
Royal Netherlands Air Force, Scheveningen, the Netherlands, June 7, 2000.
64William Drozdiak, “Air War Commander Says Kosovo Victory Near,” Washington
Post, May 24, 1999.
Friction and Operational Problems 129
that army’s ability to move and its willingness to fight, and that both
of those remained decidedly intact.65
In its final tally as Operation Allied Force ended, the U.S. Defense
Department settled on 700 out of 1,500 tanks, APCs, and artillery
pieces destroyed altogether in Kosovo.69 More specifically, General
Shelton announced in an early postwar briefing that NATO attacks
had destroyed “around 120 tanks, about 220 armored personnel car-
riers, and up to 450 artillery and mortar pieces.” However, nothing
like a matching number of hulks was found by allied inspectors after
Allied Force ended. During their withdrawal, VJ troops took hun-
dreds of tanks, artillery pieces, and APCs out of Kosovo. They also
seemed spirited and defiant rather than beaten.70 The VJ’s com-
______________
65William M. Arkin, “Limited Warfare in Kosovo Not Working,” Seattle Times, May 22,
1999.
66Paul Richter, “Milosevic War Machine Has a Lot of Fight Left,” Los Angeles Times,
April 29, 1999.
67Robert Hewson, “Allied Force, Part II: Overwhelming Air Power,” World Air Power
Journal, Winter 1999/2000, p. 113.
68 Michael Evans, “Serb Army Talks of Peace as Armor Takes a Pounding,” London
Times, June 2, 1999.
69Rowan Scarborough, “Pentagon Intends to Issue Final Count of Serbian Losses,”
Washington Times, July 9, 1999.
70Over the course of the 11-day Serb withdrawal, NATO observers counted 220 tanks,
300 APCs, and 308 artillery pieces being loaded onto trucks and transporters, along
with hundreds of other vehicles and assorted military equipment. Steven Lee Myers,
“Damage to Serb Military Less Than Expected,” New York Times, June 28, 1999.
130 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Not only did the Serbs make successful use of tank decoys made out
of tetra-pak milk carton material, they also positioned wood-burning
stoves with their chimneys angled to make them look like artillery
pieces. In some cases, water receptacles were found in the decoys,
cleverly placed there to heat up under the sun to help replicate the
infrared signature of a vehicle or hot artillery tube.74 One source
spoke of cockpit display videotapes showing targets with every ap-
pearance of being tanks collapsing instantly upon being hit. In addi-
tion, the Serbs made heavy and frequently effective use of smoke
generators to protect targets against LGBs. After the air war ended,
______________
71“Yugoslav Army Lost 524 Soldiers, Top General Says,” International Herald Tribune,
July 22, 1999.
72Briefing by Admiral James O. Ellis, USN, commander in chief, U.S. Naval Forces,
Europe, and commander, Allied Forces Southern Europe and Joint Task Force Noble
Anvil, no date given.
73Joseph Fitchett, “NATO Misjudged Bombing Damage,” International Herald Tri-
bune, June 23, 1999. General Jumper dismissed criticisms from some that expensive
U.S. precision munitions had been wasted on decoys. Declaring that U.S. forces had
had “plenty of bombs for decoys,” he noted that what appeared to be legitimate tar-
gets were immediately attacked so that aircrews would not loiter over target areas try-
ing to distinguish real targets from decoys and exposing themselves needlessly to en-
emy fire. Aviation Week and Space Technology, September 20, 1999, p. 25.
74Paul Richter, “U.S. Study of War on Yugoslavia Aimed at Boosting Performance,” Los
Angeles Times, July 10, 1999.
Friction and Operational Problems 131
The team’s initial conclusion from that assessment was that “only a
handful” of enemy tanks, APCs, and artillery pieces could be deter-
mined to have been catastrophically damaged by air attacks. 77 Al-
though the team succeeded in investigating some 60 percent of
NATO’s claimed hits on mobile targets in the KEZ, it confirmed only
14 tanks, 18 APCs, and 20 artillery pieces as destroyed for sure. A
later assessment conducted by USAFE’s office of studies and analy-
______________
75Cohen and Shelton, After-Action Report, pp. 84–85.
76As the team’s concept of operations clearly stipulated, the mission objective was to
“determine Allied Force munition effectiveness by selective examination of fixed and
mobile target sets within Kosovo [and to] evaluate and record physical and functional
target damage and precise weapons impact locations and characteristics, with em-
phasis on precision and near-precision air-dropped munitions.” The concept of oper-
ations further stipulated that validation of NATO’s air campaign, target set, BDA, and
rationale for specific target selection were “beyond the scope of this survey.” Docu-
mentation provided to the author by Hq USAFE/SA, May 2, 2001.
77 Tim Butcher and Patrick Bishop, “NATO Admits Air Campaign Failed,” London
Daily Telegraph, July 22, 1999.
132 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
These seeming discrepancies led some air war critics to charge that
NATO and the U.S. Defense Department were engaging in a blatant
cover-up of allied air power’s poor performance against VJ forces in
Kosovo to avoid being embarrassed by the paltry numbers the in-
spection team had produced. That criticism turned out, however, to
have been overblown for two reasons. First, the cover-up charge was
misdirected, in that it was based entirely on a leaked draft report by
USAFE’s inspection team that went to Kosovo earlier in the summer
of 1999. That draft report, dated August 3, 1999, and titled “Op-
eration Allied Force: Munitions Effectiveness Assessment, Vol. II:
Mobile Targets,” documented information collected in Kosovo and
elsewhere by the MEA working group tasked with looking into
mobile enemy targets. That effort was undertaken not to account for
successful strikes, but rather to determine what equipment remained
at the attacked sites. The freshest of the attacked sites visited was
four weeks old, and some were only visited for the first time three
months after the attacks.
All told, the USAFE team came across 14 tank carcasses and the hulks
of 12 self-propelled artillery vehicles, which could have looked like
tanks from the air and been reported as such in post-strike pilot
mission reports. That added up to 26 confirmable “tanks” suffering
sufficiently catastrophic damage from NATO air attacks to be written
off and abandoned by departing VJ forces. Cross-referencing pilot
reports with corroborating evidence from other sources, the USAFE
______________
78John Barry, “The Kosovo Cover-Up,” Newsweek, May 15, 2000, p. 23.
79Richard J. Newman, “The Bombs That Failed in Kosovo,” U.S. News and World Re-
port, September 20, 1999.
Friction and Operational Problems 133
______________
80Stephen P. Aubin, “Newsweek and the 14 Tanks,” Air Force Magazine, July 2000, pp.
59–61. As USAFE’s director of studies and analysis, Brigadier General John Corley,
who directed that assessment, explained afterward during a Pentagon press briefing,
“if a pilot claimed that he had attacked a tank at a given [location], we would go to that
location and . . . begin to survey that exact site. If what we had was . . . multiple
sources to confirm what had been claimed, then we would put that into a successful
strike category. Let me give you an example. If we went to one of those desired mean
points of impact and we found a bomb crater and we found shrapnel and oil down in
the bottom of that bomb crater, then we would take a digitized photo of that crater
and we would note that there would be earth scarring, as if some very heavy piece of
equipment had been dragged from that bomb crater out to a road. Then we would
compare that with both before and after imagery. You might have, for example, a
[satellite] image showing a tank in a tree line. You may go and take a look at the
cockpit video which shows that tank at that exact set of coordinates with a munition
impacting it. . . . You may then go back and discover a piece of U-2 film afterward
showing a damaged tank. You may then find out that an airborne forward air
controller who had flown specifically over this area day in and day out would report
that approximately two to three days later, whatever had been there was now gone
from that location. We further wound up with some information whereby we saw
bomb-damaged and destroyed equipment loaded on board flatbed trucks being taken
out of Kosovo, headed back north into Serbia. So as you begin to look at all those
sources of information, those multiple layers worth . . . in concert, and if we had
multiple pieces of evidentiary information, we would confirm a successful strike. And
that was the difference between the 26 and the 93. If we could not confirm with
multiple sources, we did not claim a successful strike.” News briefing, Office of the
Assistant Secretary of Defense (Public Affairs), the Pentagon, Washington, D.C., May 8,
2000.
134 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
81Indeed, in its interim report on the Kosovo air effort, the USAF expressly conceded
that “shortfalls remain . . . in the USAF’s ability to locate and attack moving armor and
other ground forces in poor weather. The Air Force needs to continue to develop and
improve its ability to do this.” The Air War Over Serbia: Aerospace Power in Operation
Allied Force, Washington, D.C., Hq United States Air Force, April 1, 2000, p. 53.
82Cohen and Shelton, After-Action Report, p. 56.
83Dana Priest, “Tension Grew with Divide in Strategy,” Washington Post, September
21, 1999.
Friction and Operational Problems 135
They’re not going to be doing anything with those forces for a long
time.”84
______________
84Ignatieff, Virtual War, p. 106.
85This is not to suggest that one should draw any particular comfort from the appar-
ent fact that NATO’s failure to take out more than a token number of VJ tanks was
largely irrelevant to the overall outcome of Allied Force. For one thing, had NATO
been able to render the VJ’s Kosovo corps ineffective during the air war’s initial month,
Milosevic may well have capitulated earlier, to the relief of both NATO and the Koso-
var Albanians. Second, and more important, the mission of finding, identifying, and
destroying dispersed and concealed enemy tanks is not going to go away, and the U.S.
Air Force will likely be asked again in some future contingency to attack fielded enemy
forces under comparably challenging circumstances. Civilians in senior leadership
positions who recall the more optimistic early claims on behalf of the air war’s accom-
plishments in this respect will naturally expect air power to perform effectively. For-
tunately, despite charges from some that the Air Force sought to play down its difficul-
ties in this regard in the early aftermath of Allied Force, its leadership has frankly
owned up to those difficulties and has initiated measures aimed at improving its ca-
pability. I am grateful to my RAND colleague Bruce Pirnie for directing my attention
to this point.
136 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
86Joel Havemann, “Convoy Deaths May Undermine Moral Authority,” Los Angeles
Times, April 15, 1999.
87Indeed, the train entered the AGM-130’s field of regard so close to the moment of
weapon impact that the F-15E weapon systems officer (WSO) who was controlling the
guiding weapon noted that he had not even seen it until the videotape of his cockpit
display was played back during the subsequent mission debriefing. As a measure of
the extent to which F-15E aircrews, like all others, were disciplined to honor the
strictest collateral-damage avoidance rules, there were numerous instances in which
the WSO dragged the selected impact point of a guiding AGM-130 off the designated
aim point to an open area at the last moment because the target looked through the
weapon’s EO seeker head like a house or some other potential opportunity for collat-
eral damage. In a similar illustration of such discipline, one videotape of an AGM-130
attack on an enemy fuel storage tank as the weapon neared impact showed the tar-
geted tank to be empty while others around it were full. Nevertheless, despite the
WSO’s natural temptation, the guiding weapon was not slewed at the last moment to-
ward a more lucrative target because the empty fuel tank happened to be the one to
which the approved DMPI had been assigned. Conversation with USAF F-15E air-
crews, 492nd Fighter Squadron, RAF Lakenheath, England, April 27, 2001.
Friction and Operational Problems 137
In the case of the Djakovica incident noted above, there were initial
reports that Yugoslav aircraft had intentionally attacked the civilian
______________
88Rowan Scarborough, “As Strikes Mount, So Do Errors,” Washington Times, May 11,
1999.
89 Robert Wall, “NATO Shifts Tactics to Attack Ground Forces,” Aviation Week and
Space Technology, April 12, 1999, p. 23.
138 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
During the course of the precision attacks with 500-lb GBU-12 LGBs
that then ensued, it was reported as “possible” that some of the ve-
hicles may have been civilian tractors, at which point the FAC-A im-
mediately called all fighters off “high and dry” (clear of the target
area with their armament switches deselected), and the ABCCC, in
turn, requested reverification of the targets as hostile. At that point,
nearby OA-10s were called in so that their pilots might reconnoiter
the situation and provide such reverification with onboard nine-
power space-stabilized binoculars. One OA-10 pilot reported observ-
ing definite military vehicles but also multicolored and possibly
civilian vehicles, whereupon the FAC-A terminated all further at-
tacks. Afterward, Serb news reports claimed that 80 civilians had
been killed, although the persistent ambiguities were such that
NATO only conceded that it “may have attacked” civilian vehicles.
Some reports suggested that the civilians involved had been
machine-gunned rather than bombed, and eyewitnesses on the
ground reported the use of human shields in the convoys and nearby
Serb mortar fire at the same time the convoy was being attacked by
the F-16s. The commander of the 31st Air Expeditionary Wing whose
F-16s were involved in the tragedy, Brigadier General Leaf, later told
______________
90Michael Dobbs and Karl Vick, “Air Strikes Kill Scores of Refugees,” Washington Post,
April 15, 1999.
Friction and Operational Problems 139
The extraordinary media attention that was given to events like these
attested to what can happen when incurring zero noncombatant ca-
sualties becomes not just the goal of strategy but also the expecta-
tion. Thanks to unrealistic efforts to treat the normal friction of war
as avoidable human error, every occurrence of unintended collateral
damage became overinflated as front-page news and treated as a
blemish on air power’s presumed ability to be consistently precise.
Indeed, the added constraints imposed on NATO aircrews as a result
of such occasional tragic occurrences indicated the degree to which
modern air power has become a victim of its own success. During
the Gulf War, cockpit video images of LGBs homing with seemingly
unerring accuracy down the air shafts of enemy bunkers were spell-
binding to most observers. Yet because of that same seemingly
unerring accuracy, such performance has since come to be expected
by both political leaders and the public alike. Once zero collateral
damage becomes accepted as a measure of strategy success, not only
air power but all forms of force employment get set up to be judged
by all but unreachably high standards. Inevitably, any collateral
damage then caused during the course of a campaign becomes grist
for domestic critics and the enemy’s propaganda mill. Anthony
Cordesman rightly noted how characterizations of modern precision
bombing as “surgical” overlook the fact that patients still die on the
operating table from time to time.92 Nevertheless, a nontrivial num-
ber of proposed sorties in Operation Allied Force were either can-
celed outright or aborted at the last minute before any weapons were
released because their targets (wryly characterized by some USAFE
staffers as “morally hardened”) could not be positively identified or
______________
91Videotaped press statement by Brigadier General Daniel Leaf, USAF, Brussels, Bel-
gium, NATO Office of Information and Press, April 19, 1999.
92Anthony H. Cordesman, “The Lessons and Non-Lessons of the Air and Missile War
in Kosovo,” unpublished draft, Center for Strategic and International Studies, Wash-
ington, D.C., July 20, 1999.
140 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
93Typical of such baseless charges was the reference by one pundit to the “low alti-
tudes at which tactical attacks work,” yet where “pilots risk getting killed” (William
Pfaff, “After NATO’s Lies About Kosovo, It’s Time to Come Clean,” International Her-
ald Tribune, May 11, 2000) and the allegation by another that “avoiding risk to pilots
multiplied the risk to civilians exponentially” (James Carroll, “The Truth About
NATO’s Air War,” Boston Globe, June 20, 2000).
94Edward N. Luttwak, “Give War a Chance,” Foreign Affairs, July/August 1999, p. 40.
Friction and Operational Problems 141
tics are employed, the timeline for target acquisition and weapons
guidance is substantially longer, thus improving the chance of
achieving a hit.
______________
95“NATO Jets May Have Erred in Convoy Attack, General Says,” Aerospace Daily, April
20, 1999, p. 102.
142 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
The point of the foregoing is that for the kinds of circumstances that
repeatedly occasioned the accidental loss of civilian life in Allied
Force, the United States, to say nothing of its NATO allies, has yet to
develop fail-safe target discrimination capabilities and tactics for use
either above or below 15,000 ft. As a result, it has had little choice but
to rely on draconian rules of engagement (ROE), which are designed
to hedge on the side of caution yet are anything but foolproof. In one
case during Allied Force in which the ROE worked as intended, a
USAF pilot was directed not to attack a confirmed SA-6 launcher be-
cause it was parked immediately adjacent to a civilian structure in a
village. There were other reported instances in which precision
munitions in the process of guiding were deliberately steered away
from targets at the last minute to avoid harming civilians who had
not been seen in the target area until after weapon release.96 In the
most egregious instance in which the ROE regime appears to have
failed, however, namely, the tragedy involving the convoy along the
Djakovica road in Kosovo, the FAC who was coordinating the attack
had been given a positive identification by the ABCCC that was
completely consistent with the prevailing ROE. Upon observing that
the vehicles were uniformly colored and evenly spaced, the FAC
declared the convoy to be a valid target. He had also been given
ABCCC approval to clear the fighters under his control to drop at will
after one F-16 orbiting overhead had drawn fire from one of the con-
voy’s vehicles.97
______________
96John A. Tirpak, “The State of Precision Engagement,” Air Force Magazine, March
2000, p. 26.
97 It further bears stressing in this regard that most cases of unintended damage
resulting in civilian deaths occurred inside targeted buildings, which were
prespecified in the ATO and against which NATO aircrews were not free to exercise
real-time discretion. Other such cases were occasioned by munitions failures such as
faulty cluster-bomb fuses or laser target designators that were disrupted by smoke or
clouds while a weapon was guiding. Neither had anything to do with weapon-release
altitude. The only clear case of noncombatant fatalities that can be even indirectly
ascribed to altitude was the April 14 Djakovica convoy incident, during which the
attack was immediately called off once the target identification error was discovered.
Friction and Operational Problems 143
______________
98Email from Lieutenant Colonel James Tubbs, AF/XPXQ, to Colonel James Callard,
AF/XPXS, February 11, 2000. Lieutenant Colonel Tubbs was the operations officer of
the 510th Fighter Squadron flying F-16CGs out of Aviano Air Base during Operation
Allied Force.
99Although, as in Desert Storm, AWACS generally provided a superb threat picture to
allied pilots operating in hostile airspace, at least one specific instance of friction was
reported by a USAF F-15C pilot who downed a Yugoslav MiG-29 during a day defen-
sive counterair mission on March 26. The pilot complained that the supporting
AWACS controller “did not have any inkling [that] someone was flying on the other
side of the border, although he was real good at calling out every friendly west of us”
(email communication to the author, June 4, 1999). The F-15 pilot further charged
that the supporting AWACS was still unaware of the MiG-29’s presence even after ini-
tial moves had commenced. The intercepting pilot accordingly assessed the assumed
threat aircraft to be hostile by origin, since there were no NATO offensive counterair
missions airborne at the time. Only after the engagement was fully joined and the
F-15 pilot had visually confirmed his target to be a MiG-29 did the AWACS controller
finally report two possible hostile contacts in lead-trail formation.
144 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
100 The error was also reminiscent of earlier damage to the French embassy in Tripoli,
Libya, in 1986 during the joint U.S. Air Force–U.S. Navy Operation El Dorado Canyon
against Libya’s ruler, Moammar Khaddafi, caused when the bomb fragmentation pat-
tern from a preceding F-111 forced the trailing pilot to shift course, inadvertently
sending his bombs into the embassy. That, however, was an operational error occa-
sioned by the heat of battle, not a planning error committed by target nominators.
Friction and Operational Problems 145
As it turned out, U.S. intelligence had the correct street address for
the intended target, which was a Yugoslav weapons-producing
agency called the Federal Directorate of Supply and Procurement.
______________
101 David A. Fulghum and Robert Wall, “Intel Mistakes Trigger Chinese Embassy
Bombing,” Aviation Week and Space Technology, May 17, 1999, p. 55.
102 Eric Schmitt, “Aim, Not Arms, at the Root of Mistaken Strike on Embassy,” New
York Times, May 10, 1999.
103 Paul Richter and Doyle McManus, “Pentagon to Tighten Targeting Procedures,”
Los Angeles Times, May 11, 1999.
146 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
104 Vernon Loeb and Steven Mufson, “CIA Analyst Raised Alert on China’s Embassy,”
Washington Post, June 24, 1999.
105 Eric Schmitt, “Pentagon Admits Its Maps of Belgrade Are Out of Date,” New York
Times, May 11, 1999, and Bradley Graham, “U.S. Analysts Misread, Relied on Outdated
Maps,” Washington Post, May 11, 1999.
106 Steven Pearlstein, “NATO Bomb Said to Hit Belgrade Hospital,” Washington Post,
May 21, 1999.
Friction and Operational Problems 147
______________
107 Steven Lee Myers, “Chinese Embassy Bombing: A Wide Net of Blame,” New York
Times, April 17, 2000.
108 Ivo H. Daalder and Michael E. O’Hanlon, Winning Ugly: NATO’s War to Save
Kosovo, Washington, D.C., The Brookings Institution, 2000, p. 147.
148 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
109 Another reported problem with the Macedonia basing option was the fact that it
would have been a violation of the Dayton accords to station any offensive forces
within the territorial confines of the former Yugoslavia. Albania was thus the only real-
istic alternative.
110 Bradley Graham and Dana Priest, “Allies to Begin Flying Refugees Abroad,” Wash-
ington Post, April 5, 1999.
Friction and Operational Problems 149
Accordingly, the Army was driven by its own standard operating pro-
cedures to supplement the two Apache battalions with an additional
heavy contingent of ground forces, air defenses, military engineers,
and headquarters overhead. As the core of this larger force comple-
ment, now designated Task Force (TF) Hawk, the Apaches were
drawn from the Army’s 11th Aviation Brigade stationed at Illesheim,
Germany. The deployment package included, however, not only the
two battalions of AH-64s, but also 26 UH-60L Blackhawk and
CH-47D Chinook helicopters from the 12th Aviation Regiment at
Wiesbaden, Germany. Additional assets whose deployment was
deemed essential for supporting the Apaches included a light in-
fantry company; a multiple-launch rocket system (MLRS) platoon
with three MLRS vehicles; a high-mobility multipurpose wheeled
vehicle (HMMWV, or “humvee”) antitank company equipped with 38
armed utility vehicles; a military intelligence platoon; a military po-
lice platoon; and a combat service support team. The Army further
determined a need for its Apaches to be accompanied by a mecha-
nized infantry company equipped with 14 Bradley AFVs; an armor
company with 15 M1A2 Abrams main battle tanks; a howitzer battery
with eight 155mm artillery pieces; a construction engineer company;
a short-range air defense battery with eight more Bradley AFVs
armed with Stinger infrared SAMs; a smoke generator platoon; a
brigade headquarters complement; and diverse other elements. In
all, to backstop the deployment of 24 attack helicopters to Albania,
TF Hawk ended up being accompanied by a support train of no fewer
than 5,350 Army personnel.
In all events, the Apaches with their attached equipment and per-
sonnel arrived in Albania in late April. No sooner had the Army de-
clared all but one of the aircraft ready for combat on April 26 when,
only hours later, one crashed at the Tirana airfield in full view of re-
porters who had been authorized to televise the flight. (The 24th
Apache had developed hydraulic trouble en route and remained on
the ground in Italy.) Neither crewmember was injured, but the acci-
dent was an inauspicious start for the widely touted deployment.
Less than two weeks later, on May 5, a second accident occurred, this
time killing both crewmembers during a night training mission some
46 miles north of Tirana. The aircraft was carrying a full load of
______________
111 That said, it bears noting that the threat of Serbian forces coming across the Alba-
nian border did not appear to be a matter of great concern to anyone in the Allied
Force command hierarchy before the arrival of TF Hawk, even though there were U.S.
troops already on the ground in Albania as a part of JTF Shining Hope, the Albanian
refugee relief effort, who were not provided with any comparable force-protection
package.
112 Elaine M. Grossman, “Army’s Cold War Orientation Slowed Apache Deployment to
Balkans,” Inside the Pentagon, May 6, 1999, p. 6. Notably, the C-17 demonstrated for
the first time the ability to air-deliver a significant Army force of M1 tanks, M2 AFVs,
MLRSs, howitzers, and engineering equipment.
Friction and Operational Problems 151
As of May 31, the cost of the TF Hawk deployment had reached $254
million, much of that constituting the expense for the hundreds of
C-17 sorties that had been needed to haul all the equipment from
Germany to Albania, plus the additional costs of building base camps
and port services and conducting mission rehearsals.114 Yet despite
SACEUR’s intentions to the contrary, the Apaches flew not a single
combat mission during the entire remainder of Operation Allied
Force. The reason given afterward by the JCS chairman, General
Shelton, was that Serb air defenses in Kosovo, although noticeably
degraded by early May, remained effective enough to warrant keep-
ing the Apaches out of action until SEAD operations had “reduced
the risk to the very minimum.”115
______________
113 Paul Richter and Lisa Getter, “Mechanical Error, Pilot Error Led to Apache
Crashes,” Los Angeles Times, May 13, 1999.
114 Ron Lorenzo, “Apache Deployment Has Cost Quarter Billion So Far,” Defense
Week, June 7, 1999, p. 6.
115 Molly Moore and Bradley Graham, “NATO Plans for Peace, Not Ground Invasion,”
Washington Post, May 17, 1999.
116 Sheila Foote, “Shelton: Risk Was the Key in Decision Not to Use Apaches,” Defense
Daily, September 10, 1999, p. 2.
152 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
incurring less risk from enemy infrared SAMs, AAA, and small-arms
fire than the Apaches would have faced.117
Beyond the problems created for the deployment by the Army’s deci-
sion to bring along so much additional overhead, there was a break-
down in joint doctrine for the combat use of the helicopters that was
disturbingly evocative of the earlier competition for ownership and
control of coalition air assets that had continually poisoned the rela-
tionship between the joint force air component commander (JFACC)
and the Army’s corps commanders during Desert Storm.118 The is-
sue stemmed in this case from the fact that the Army has tradition-
ally regarded its attack helicopters not as part of a larger air power
equation with a theater-wide focus, but rather as an organic maneu-
ver element fielded to help support the ground maneuver needs of a
division or corps. Apache crews typically rely on their own ground
units to select and designate their targets. Yet in the case of Allied
Force, with no Army ground combat presence in theater to speak of,
they would either have had to self-designate their targets or else rely
on Air Force forward air controllers flying at higher altitudes to des-
ignate for them. The idea of using Apaches as a strike asset in this
manner independently of U.S. ground forces was simply not recog-
nized by prevailing Army doctrine. On the contrary, as prescribed in
Army Field Manual FM 1-112, Attack Helicopter Operations, an AH-64
battalion “never fights alone. . . . Attacks may be conducted out of
physical contact with other friendly forces,” but they must be
“synchronized with their scheme of maneuver.” FM 1-112 expressly
characterizes deep-attack missions of the sort envisaged by Clark as
“high-risk, high-payoff operations that must be exercised with the
utmost care.”119
______________
117 True enough, the terrain and weather presented by Kosovo were more challenging
than the open and featureless Iraqi desert, where the Apaches had performed so ef-
fectively against enemy armor in Desert Storm. Yet the biggest concern in the minds
of many U.S. leaders was the specter of a replay of the 1993 “Bloody Sunday” horror in
Mogadishu, Somalia, with dead Army Rangers and crewmembers from downed Black-
hawk helicopters being dragged through the streets on live television worldwide.
118 David Atkinson and Hunter Keeter, “Apache Role in Kosovo Illustrates Cracks in
Joint Doctrine,” Defense Daily, May 26, 1999, p. 6.
119 Quoted in Elaine M. Grossman, “As Apaches Near Combat, White House Seeks
Diplomatic Solution,” Inside the Pentagon, May 6, 1999, p. 7.
Friction and Operational Problems 153
______________
120 Telephone conversation with Lieutenant General Michael Short, USAF (Ret.),
August 22, 2001.
154 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
(DOCC) reportedly put it, “they do not know, nor do they want to
know, the detailed integration required to get the Prowler to jam the
priority threats, provide acquisition jamming on the correct azimuth,
etc. . . . The benefits of integrating with platforms like Compass Call,
Rivet Joint and others are off their radar scope.”121
After Allied Force ended, the assistant chief of staff for operations at
Supreme Headquarters Allied Power Europe (SHAPE), USAF Major
General John Dallager, touched the heart of the overriding interests
and equities at stake here when he stated, during a briefing at a
NATO Reaction Force Air Staff conference on JFACC issues: “Clearly
the JFACC’s authority must not infringe upon operational C2
[command and control] relationships within and between national
or service commands and other functional commands. But to
ensure deconfliction of simultaneous missions and to minimize the
risk of fratricide, all air operations within the [joint operating arena]
must be closely coordinated by the JFACC through the ATO . . .
process. This last point may be difficult to swallow for land and
maritime commanders, but if air history teaches us anything, it is
that air, the truly joint activity, needs to be coordinated centrally if
we are to make efficient use of scarce resources and if we are to avoid
blue-on-blue.”122
______________
121 Elaine M. Grossman, “Army Commander in Albania Resists Joint Control over
Apache Missions,” Inside the Pentagon, May 20, 1999, p. 9. In his memoirs, Clark later
scored this article for “personally attacking Jay Hendrix and claiming, among other ac-
cusations, that he would not allow the Apache sorties to appear on Short’s Air Tasking
Order.” Clark made no attempt to refute that accusation, however, but merely dis-
missed it as the complaint of a “disgruntled Air Force officer” whose “misunder-
standing, communicated without perspective to friends in other units, suddenly
surfaced to make news weeks after it had been written, after the problems it ad-
dressed, if real then, had been corrected.” General Wesley K. Clark, Waging Modern
War: Bosnia, Kosovo, and the Future of Combat, New York, Public Affairs, 2001, p. 320.
122 Major General John Dallager, USAF, “NATO JFACC Doctrine,” briefing at a confer-
ence on “The NATO Joint Force Air Component Commander Concept in Light of the
Kosovo Air Campaign,” Headquarters NATO Reaction Force Air Staff, Kalkar, Ger-
many, December 1–3, 1999. It might be noted in passing here that another Army–Air
Force difference of view that had an even greater operational impact than the joint
doctrinal disagreement discussed above (because all involved had to live through its
consequences) was the disconnect between the two services at Tirana as to who was in
charge of the airfield and force protection, a disconnect that, according to one senior
USAF planner who was involved, created “some real problems.” Comments on an
earlier draft by Brigadier General Robert Bishop, Hq USAF/XOO, April 17, 2001.
Friction and Operational Problems 155
______________
123 Ibid.
156 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
The TF Hawk experience underscored how little the U.S. Army, by its
own leadership’s candid admission, had done since Desert Storm to
increase its capacity to get to an emergent theater of operations
rapidly and with sufficient forces to offer a credible combat presence.
Shortly after the Gulf War, the Army’s leadership for a time enter-
tained the thought of reorganizing the service so that it might be-
come more agile by abandoning its structure of 10 combat divisions
and opting instead for 25 “mobile combat groups” of around 5,000
troops each. Ultimately, however, the Army backed away from that
proposed reform, doing itself out of any ability to deploy a strong
armored force rapidly and retaining the unpalatable alternatives of
airlifting several thousand lightly armed infantrymen to a theater of
conflict within days or shipping a contingent of 70-ton M1A2 Abrams
main battle tanks over the course of several months.125
On his second day in office as the Army’s new chief of staff, General
Shinseki acknowledged that the Army had been poorly prepared to
move its Apaches and support overhead to Albania. Part of the
problem, he noted fairly, was that the only available deployment site
that made any operational sense had poor rail connections, a shallow
port, and a limited airfield capacity that could not accommodate the
Air Force’s C-5 heavy airlifter. However, he admitted that the Army
______________
124 George C. Wilson, “Memo Says Apaches, Pilots Were Not Ready,” European Stars
and Stripes, June 20, 1999.
125 Thomas E. Ricks, “Why the U.S. Army Is Ill-Equipped to Move Troops Quickly into
Kosovo,” Wall Street Journal, April 16, 1999. The most fully developed and widely
cited articulation of this proposed Army reorganization, which failed to take root, may
be found in Colonel Douglas A. Macgregor, USA, Breaking the Phalanx: A New Design
for Landpower in the 21st Century, Westport, Connecticut, Praeger, 1997.
Friction and Operational Problems 157
______________
126 Eric Schmitt, “New Army Chief Seeks More Agility and Power,” New York Times,
June 24, 1999.
127 “Shinseki Hints at Restructuring, Aggressive Changes for the Army,” Inside the
Army, June 28, 1999, p. 1.
128 Lieutenant General Theodore G. Stroup, Jr., USA (Ret.), “Task Force Hawk: Beyond
Expectations,” Army Magazine, August 1999.
158 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
129 Response to a question at an Air Force Association Eaker Institute colloquy,
“Operation Allied Force: Strategy, Execution, Implications,” held at the Ronald
Reagan International Trade Center, Washington, D.C., August 16, 1999.
130 Comments on an earlier draft by Brigadier General Robert Bishop, Hq USAF/XOO,
April 12, 2001, and Colonel Robert Owen, Hq AMC, May 10, 2001. See also General
Charles T. Robertson, Jr., USAF, commander in chief, U.S. Transportation Command,
and commander, Air Mobility Command, “Air War Over Serbia: A Mobility
Perspective,” briefing charts, 2000, Hq USAFE/SA library.
Friction and Operational Problems 159
RAND MR1365-6.3
1600
C-17
1400
C-130
C-141
1200
C-5
Short tons delivered
1000 Other
800
600
400
200
0
24
31
14
21
28
12
19
26
9
ril
ay
ne
ne
ch
ch
ril
ril
ril
ay
ay
ay
Ap
Ju
Ju
Ap
Ap
Ap
M
ar
ar
M
SOURCE: Hq USAFE/SA.
______________
131 The AGM-130 is a rocket-boosted variant of the electro-optical and infrared guided
GBU-15 2,000-lb PGM featuring midcourse GPS guidance updates. At the start of the
air war, 200 of these weapons had been fielded, and those used were pulled from Air
Combat Command’s Weapons System Evaluation Program (WSEP), leaving no muni-
tions for training. William M. Arkin, “Kosovo Report Short on Weapons Performance
Details,” Defense Daily, February 10, 2000, p. 2.
132 Ibid.
133 Lieutenant General Marvin R. Esmond, testimony to the Military Procurement
Subcommittee, House Armed Services Committee, Washington, D.C., October 19,
1999.
134 David A. Fulghum, “Lessons Learned May Be Flawed,” Aviation Week and Space
Technology, June 14, 1999, p. 205.
Friction and Operational Problems 161
without the aid of the Internet, which made for vastly more rapid
worldwide information availability than did the former hard-copy
practices. Frequently, however, because of the absence of institu-
tionalized procedures, the use of SIPRNET made for confusion and
difficulty in finding some target materials on short notice. In addi-
tion, real-time target information would be withheld from U.S. allies
as U.S. officials argued over who should be allowed to see what. Fi-
nally, NIMA was frequently slow to deliver overhead photography of
proposed targets and of targets already attacked, which in turn
slowed the battle-damage assessment process and the decision as to
whether to retarget a previously attacked site. One informed source
commented that ISR fusion worked better in Allied Force than it did
during Desert Storm, but that it still rated, at best, only a grade of
C-plus in light of what remained to be done. In contrast, what gener-
ally worked well was the “reach-back” procedure first pioneered in
Desert Storm, in which commanders and planners in the forward
theater used secure communications lines to tap into information
sources in the intelligence community in Washington and else-
where. 135
______________
135 Rowan Scarborough, “Kosovo Target Data Stalled in Transit,” Washington Times,
July 28, 1999.
136 Tim Ripley, “Tanker Operations,” World Air Power Journal, Winter 1999/2000,
p. 121.
162 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
137 Wall, “Airspace Control Challenges Allies.”
Friction and Operational Problems 163
______________
138 Colonel E. Baldazzi, Italian Air Force, “Host Nation Support for the Kosovo Air
Campaign,” briefing at a conference on “The NATO Joint Force Air Component Com-
mander Concept in Light of the Kosovo Air Campaign,” Headquarters NATO Reaction
Force Air Staff, Kalkar, Germany, December 1–3, 1999.
139 Joseph Fitchett, “For NATO, Keeping Peak Air Traffic on the Go Was a Critical
Goal,” International Herald Tribune, March 31, 2000.
164 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
lash against the effort. Another was to avoid any replay of the
downing of an Iranian airliner, which the cruiser USS Vincennes
mistook for an Iranian F-14 over the Persian Gulf in 1988. That latter
concern led to a double-checking of identification procedures for
electronically identifying aircraft operating in and near the combat
zone. Toward the end of the air war, NATO finally succeeded in eas-
ing the airspace congestion problem at least marginally, when it in
effect opened a second front by initiating Marine F/A-18D opera-
tions out of Hungary and USAF fighter operations out of Turkey.
______________
140 “Space Support to Operation Allied Force: Preliminary Lessons Learned,” briefing
to the author by Colonel Robert Bivins, Director of Operations, U.S. Air Force Space
Warfare Center, Schriever AFB, Colorado, February 25, 2000.
Friction and Operational Problems 165
______________
141 Bill Gertz and Rowan Scarborough, “Dangerous Drawdown,” Washington Times,
April 30, 1999.
142 “Space Support to Operation Allied Force: Preliminary Lessons Learned,” briefing
to the author by Colonel Robert Bivins, director of operations, U.S. Air Force Space
Warfare Center, Schriever AFB, Colorado, February 25, 2000.
143 Roy Bender, “Allies Still Lack Real-Time Retargeting,” Jane’s Defense Weekly, April
7, 1999.
166 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
INTEROPERABILITY PROBLEMS
One of the most surprising aspects of the Allied Force experience was
what it revealed about the extent of the discontinuity that had been
allowed to develop between U.S air power and that of most other
NATO allies who participated. 146 One concern had to do with inade-
______________
144 John Donnelly, “NRO Chief: Services Ill-Prepared to Work with Spy Satellites,”
Defense Week, July 12, 1999, p. 2.
145 Quoted in The Air War Over Serbia, p. 53.
146 For a fuller treatment of the allied contribution to the air war and the interoper-
ability problems that became manifest as a result of it, see John E. Peters, Stuart John-
Friction and Operational Problems 167
_____________________________________________________________
son, Nora Bensahel, Timothy Liston, and Traci Williams, European Contributions to
Operation Allied Force: Implications for Transatlantic Cooperation, Santa Monica,
California, RAND, MR-1391-AF, 2001.
147 Since allied aircraft could not receive Have Quick radio transmissions and since
enemy forces made no effort to jam allied UHF communications, which Have Quick
was expressly developed to counter, the Have Quick capability was not used by U.S.
combat aircrews during Allied Force.
148 “NATO Jets May Have Erred in Convoy Attack, General Says,” p. 102.
149 David A. Fulghum and Robert Wall, “Data Link, EW Problems Pinpointed by Pen-
tagon,” Aviation Week and Space Technology, September 6, 1999, pp. 87–88. The JTIDS
offers aircrews a planform view of their tactical situation, as well as a capability for
real-time exchange of digital information between aircraft on relative positions,
weapons availability, and fuel states, among other things. It further shows the posi-
tion of all aircraft in a formation, as well as the location of enemy aircraft and ground
threats. Fighters can receive this information passively, without highlighting them-
168 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Second, among all participating allied air forces, only U.S., British,
Canadian, French, Spanish, and Dutch combat aircraft had the abil-
ity to deliver LGBs without offboard target designation assistance.
General Short frankly admitted that he could not risk sending the air-
craft of many allied countries into harm’s way because of concern for
the safety of their pilots and for the civilian casualties that might be
caused by inaccurately aimed weapons. Largely for that reason,
around 80 percent of all strike sorties flown in Allied Force were car-
ried out by U.S. aircraft.
To be sure, not all participating allied air forces suffered equally pro-
nounced problems with respect to capability and versatility. The
Royal Netherlands Air Force, for example, not only kept its F-16s up
to date but also provided some aerial refueling capability. The Dutch
_____________________________________________________________
selves through radio voice communications. See William B. Scott, “JTIDS Provides
F-15Cs ‘God’s Eye View,’” Aviation Week and Space Technology, April 29, 1996, p. 63.
150 John D. Morrocco, “Kosovo Reveals NATO Interoperability Woes,” Aviation Week
and Space Technology, August 9, 1999, p. 32.
151 Barton Gellman and William Drozdiak, “Conflict Halts Momentum for Broader
Agenda,” Washington Post, June 6, 1999.
Friction and Operational Problems 169
With the USAF now out of the manned tactical reconnaissance busi-
ness altogether and the Navy’s TARPS-equipped F-14s providing the
only remaining U.S. operational capability of that nature, three of the
five remaining French Mirage IVP supersonic bombers, since con-
verted to the reconnaissance role, added valuable support by being
flown daily when the weather permitted, accounting in the end for 20
percent of the Allied Force reconnaissance missions. Operating out
of Solenzara, Italy, they flew at 40,000–50,000 ft at a speed of Mach
2.05, typically entering the war zone over Belgrade and exiting over
Kosovo, covering some 20 targets on each flight in around 15 min-
utes. Returning traditional wet-film photographs to Solenzara, they
eventually developed a routine whereby high-quality images anno-
______________
152 Conversation with Major General P. J. M. Godderij, deputy commander in chief,
RNLAF, Scheveningen, the Netherlands, June 7, 2000.
170 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
153 Chris Pocock, “Mirage IV Reconnaissance Missions,” World Air Power Journal,
Winter 1999/2000, p. 111.
154 Rowan Scarborough, “Record Deployments Take Toll on Military,” Washington
Times, March 28, 2000.
155 Hewson, “Operation Allied Force,” p. 21.
Friction and Operational Problems 171
______________
156 The principal deputy assistant secretary of the Air Force for acquisition at the time,
then–Lieutenant General Gregory Martin, acknowledged that the shortage of JDAMs
was the result of a conscious choice made five years ago to emphasize other procure-
ment needs. David A. Fulghum, “Bomb Shortage Was No Mistake,” Aviation Week and
Space Technology, May 17, 1999, p. 55.
157 See Rowan Scarborough, “Smaller U.S. Military Is Spread Thin,” Washington
Times, March 31, 1999.
158 Ibid.
159 Bradley Graham, “General Says U.S. Readiness Is Ailing,” Washington Post, April
30, 1999.
172 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
A second indication of the extent to which the U.S. military had come
to find itself strapped as a result of the force drawdown was the
sharply increased personnel tempo that was set in motion by the air
effort. In all, some 40 percent of the active-duty U.S. Air Force was
committed to Operation Allied Force and to the concurrent Opera-
tions Northern and Southern Watch over Iraq. That was roughly the
same percentage of Air Force personnel that had been committed
during Operation Desert Storm, when the total force was much
larger. Among other things, as noted earlier in Chapter Three, the
heightened personnel tempo obliged President Clinton to approve a
Presidential Selected Reserve Call-Up authorizing a summons of up
to 33,102 selected reservists to active duty. 162 It further prompted
the Air Force chief of staff, General Michael Ryan, to insist that the
USAF needed a recovery time no less than that routinely granted to
the Navy every time one of its carriers returns from a deployment.
Ryan flatly declared that “we are not a two-MTW [major theater war]
Air Force in a lot of areas, and one of them is airlift.” That shortfall
made for one of many reasons why the Air Force later insisted that it
needs 90 days to reconstitute its forces between MTWs.163
Earlier, as Allied Force entered its second month, Ryan told reporters
that “the U.S. Air Force is in a major theater war.” (He later amended
that remark to indicate that he had meant to say that the Air Force’s
commitment level included Operations Northern and Southern
______________
160 Ibid.
161 John A. Tirpak, “The First Six Weeks,” Air Force Magazine, June 1999, p. 27.
162 “U.S. Mobilizes Guard, Reserve for Balkan Duty,” Air Force Magazine, June 1999,
p. 16.
163 Vince Crawley, “Air Force Needs 90 Days Between Wars, Chief Says,” Defense Week,
August 9, 1999, p. 12.
Friction and Operational Problems 173
Watch over Iraq.)164 In the eight years since Desert Storm, deploy-
ment demands on Air Force assets had never before exceeded the
level of two AEFs of around 175 aircraft each. NATO’s air war for
Kosovo, however, demanded four AEF-equivalents’ worth of USAF
assets. Then-acting Secretary of the Air Force F. Whitten Peters de-
clared that as a result, the AEF concept would need to be reexam-
ined.165
Third, the demands of Allied Force placed a severe strain on such low
density/high demand (LD/HD) aircraft as Joint STARS, AWACS, the
U-2, the B-2, the F-16CJ, and the EA-6B.166 So many of these scarce
assets were committed to the air effort that day-to-day training in
home units suffered major shortfalls as a result. The most acute
strains were felt in the areas of surveillance, SEAD, and combat
search and rescue. Almost every Block 50 F-16CJ in line service was
committed to support SEAD operations, necessitating a virtual halt
to mission employment training in the United States. (Figure 6.4
shows the overall USAF commitment to Allied Force, broken down
by aircraft type.)
______________
164 Tirpak, “The First Six Weeks,” p. 27.
165 John T. Correll, “Assumptions Fall in Kosovo,” Air Force Magazine, June 1999, p. 4.
166 Characteristics of LD/HD include single-unit asset, limited numbers of aircraft and
pilots, and likely tasking in more than one theater. Joint Vision 2010, the “revolution
in military affairs,” improved sensor to shooter links, and decisive attack operations all
depend on more support to LD/HD assets. They transcend individual service and
weapon system boundaries.
167 Dale Eisman, “Kosovo Lesson: Navy Says It Needs More High-Tech Tools,” Norfolk
Virginian-Pilot, June 10, 1999.
174 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
RAND MR1365-6.4
Total: 517
Intratheater
airlift A-10
(43) (40)
F-15C
Special operations/ (18)
rescue/other
(38)
F-15E
(32)
ISR*
(25)
F-16CG
(35)
KC-10
(24)
F-16CJ
(64)
KC-135
(151) F-117
(25)
B-52
B-2 B-1 (11)
(6) (5)
*ISR includes RQ-1, E-3, E-8, RC-135, U-2, and EC-130 ABCCC.
SOURCE: AWOS Fact Sheet.
An even greater demand was imposed on the Air Force’s various ISR
platforms, which left none available for day-to-day continuation
training once the needs of Allied Force were superimposed on preex-
isting commitments. During the time in question, the Air Force had
______________
168 Greg Seigle, “Prowler Jammers Used to Aid NATO Air Assault,” Jane’s Defense
Weekly, March 31, 1999.
Friction and Operational Problems 175
only four operational E-8 Joint STARS aircraft, two of which were
committed to Allied Force (it has since acquired a fifth). As a result,
the Joint STARS community found itself so stripped of its most skilled
personnel that there was no instructor cadre left to work with new
crewmembers who were undergoing conversion training. The low
Joint STARS availability rate made for a typical Allied Force E-8 mis-
sion length of more than 17 hours, with the longest missions lasting
21 hours. It took two or more inflight refuelings and backup pilots
and crews to sustain each mission.169 Some Joint STARS aircraft
were flown at more than three times their normal use rates, creating
a major maintenance and depot backlog that would take months to
clear up. In all, U.S. LD/HD assets were stretched to their limit with
tasking demands whose reverberations will continue to be felt for
years in the areas of platforms, systems, reliability, parts, personnel,
retention, and replacement costs. On this point, Admiral Ellis cau-
tioned that the trend line is working in precisely the wrong direc-
tion—the demand for these assets in the future will only grow and
they should be viewed as national assets requiring joint funding, irre-
spective of service, as the highest priority. 170
______________
169 Edmund L. Andrews, “Aboard Advanced Radar Flight, U.S. Watches Combat
Zone,” New York Times, June 14, 1999.
170 Briefing by Admiral James O. Ellis, USN, commander in chief, U.S. Naval Forces,
Europe, and commander, Allied Forces Southern Europe and Joint Task Force Noble
Anvil, “The View from the Top,” 1999.
176 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
The Air Force was similarly forced to juggle scarce assets to handle
the overlapping demands imposed by Kosovo, Iraq, and Korea. It
positively scrambled to find enough tankers to support NATO mis-
sion needs in Allied Force. Ironically, both Kosovo and Iraq, in and
of themselves, represented lesser contingencies whose accommoda-
tion was not supposed to impede the U.S. military’s ability to handle
two major theater wars. Yet the burdens of both began to raise seri-
ous doubts as to whether the two-MTW construct, at least at its cur-
rent funding level, was realistic for U.S. needs. For example, when
USEUCOM redeployed 10 F-15s and 3 EA-6Bs from Incirlik to sup-
port Clark’s requirements for Allied Force, it was forced to suspend
its air patrols over northern Iraq immediately. Air patrols to enforce
the no-fly zone over southern Iraq were continued, but at a slower
operational tempo. The net result was U.S. aircraft being flown two
to three times more often than in normal peacetime operations.171
______________
171 Elizabeth Becker, “Needed on Several Fronts, U.S. Jet Force Is Strained,” New York
Times, April 6, 1999.
Friction and Operational Problems 177
fires at a much lower power and rate, with the result that the noted
effects were not discovered until they were actually seen in combat—
usually in the middle of a drop.172 Bowing to the inevitable, General
Shelton finally acknowledged the cumulative impact of these
multiple untoward trends when he admitted to Congress at the
beginning of May 1999 that there was “anecdotal and now mea-
surable evidence . . . that our current readiness is fraying and that the
long-term health of the total force is in jeopardy.”173
______________
172 Comments on an earlier draft by Hq USAF/XOXS, July 9, 1999.
173 Kate O’Beirne, “Defenseless: The Military’s Hollow Ring,” National Review, May 3,
1999, p. 18.
This page intentionally blank
Chapter Seven
LAPSES IN STRATEGY AND IMPLEMENTATION
Indeed, the six years that preceded Allied Force saw a clear regres-
sion in the use of air power after the latter’s casebook performance in
Desert Storm. With the singular exception of Operation Deliberate
Force in 1995, a trend toward what came to be called “cruise missile
diplomacy” had instead become the prevailing U.S. pattern, owing to
the ability of cruise missiles to deliver a punitive message without
risking the lives of any U.S. aircrews. The origins of this pattern went
back to June 1993, when President Clinton first ordered the firing of
several TLAMs in the dead of night against an empty governmental
building in Baghdad in symbolic reprisal for confirmed evidence that
Saddam Hussein had underwritten an assassination attempt against
former President George Bush.
______________
1Paul Richter, “U.S. Study of War on Yugoslavia Aimed at Boosting Performance,” Los
Angeles Times, July 10, 1999.
179
180 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
That may have been the administration’s going-in hope for Opera-
tion Allied Force as well. Not long after the effort began, however,
senior U.S. military leaders began voicing off-the-record misgivings
over the slow pace of the air operation, its restricted target base, and
its rules of engagement that all but proscribed any serious applica-
tion of air power. One Air Force general spoke of officers in Europe
who had characterized the air war to date as “a disgrace,” adding that
“senior military officers think that the tempo is so disgustingly slow it
______________
2In fairness to the Clinton administration, it must be said that bombing the Bosnian
Serbs unilaterally was not a realistic option for the United States as long as three
NATO allies (France, the Netherlands, and the United Kingdom) had troops on the
ground who would have been helpless against Serb reprisals had U.S. air strikes taken
place. It was only after the UN Protection Force (UNPROFOR) pulled back into de-
fensible positions so that the Serbs could not take its troops hostage that Operation
Deliberate Force became politically feasible. Weakness on the ground can often
negate strength in the air.
3For an informed, if also sharply judgmental, account of this history, see Joshua Mu-
ravchik, “The Road to Kosovo, Commentary, June 1999, pp. 17–23. See also Lieutenant
Colonel Paul K. White, USAF, Crises After the Storm: An Appraisal of U.S. Air Opera-
tions in Iraq Since the Persian Gulf War, Military Research Papers No. 2, Washington,
D.C., Washington Institute for Near East Policy, 1999.
Lapses in Strategy and Implementation 181
______________
4 Rowan Scarborough, “U.S. Pilots Call NATO Targeting a ‘Disgrace,’” Washington
Times, April 1, 1999.
5 John D. Morrocco, David Fulghum, and Robert Wall, “Weather, Weapons Dearth
Slow NATO Strikes,” Aviation Week and Space Technology, April 5, 1999, p. 26.
6William M. Arkin, “Inside the Air Force, Officers Are Frustrated About the Air War,”
Washington Post, April 25, 1999.
7To illustrate, Clark recalled after the cease-fire that he would often have to call Solana
at the last minute with an urgent request like: “You’ve got to help me with target 183.
I need 183.” Michael Ignatieff, “The Virtual Commander: How NATO Invented a New
Kind of War,” The New Yorker, August 2, 1999, p. 34.
8Ibid.
182 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
were said by some to be locked into a resigned “we can’t do it” posi-
tion rather than amenable to a more creative “let’s try it” attitude.9
______________
9Roundtable discussion with Hq USAFE/XP, USAFE/DO, and USAFE/IN staff, Ram-
stein AB, Germany, May 2, 2001.
Lapses in Strategy and Implementation 183
One reason for NATO’s overconfidence that air power alone would
suffice in forcing Milosevic to yield on Kosovo was almost surely a
misreading of the earlier Bosnian war and the role of Operation De-
liberate Force in producing the Dayton accords of 1995. As has been
widely noted since Allied Force ended, Bosnia was a part of the for-
mer Yugoslav Federation where Milosevic generally got what he
wanted and to which he was not particularly deeply attached. In the
negotiations that eventually yielded the Dayton accords, Milosevic
succeeded in keeping Kosovo unburdened by their strictures at the
price of abandoning Sarajevo to the Muslims, in a direct and outright
betrayal of his Bosnian Serb compatriots, because there was no sig-
nificant Serb minority living there.
______________
10Michael Ignatieff, Virtual War: Kosovo and Beyond, New York, Henry Holt and
Company, Inc., 2000, p. 24.
11 As one observer wrote of Operation Allied Force afterward, “so low was NATO’s
credibility with Milosevic that the threat of war and even war itself were not enough to
convince him that he had anything to fear.” Christopher Cviic, “A Victory All the
Same,” Survival, Summer 2000, p. 178.
184 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
including the mass expulsion of the Serbs from the Croatian Krajina.
Also, the 1995 bombing was not against Serbia proper, and thus did
not arouse the same nationalist response as would the bombing in
1999. The real lesson of those 1995 events might be a very different
one: that if NATO wants to have some effect, including through air
power, it needs to have allies among the local belligerents and a
credible land-force component to its strategy.”12 A false assumption
that air power alone had produced the Dayton accords may thus
have contributed further to NATO’s miscalculation that Milosevic
could be induced to give up in Kosovo after merely a few days of to-
ken bombing.13 Aleksa Djilas, son of the Yugoslav cold-war dissident
Milovan Djilas and an able intellectual in his own right, attested from
first-hand knowledge that the West had “badly underestimated the
Serbian attachment to Kosovo.” 14 In light of that, rather than ask
why it took so long for NATO’s bombing to coerce Milosevic to back
down, a more appropriate question might be why he yielded as
quickly as he did.
______________
12Adam Roberts, “NATO’s ‘Humanitarian War’ over Kosovo,” Survival, Autumn 1999,
pp. 110–111.
13This point bears emphasizing. It was not just that Serbia’s stakes in Kosovo were
much higher than in Bosnia. The two cases diverged additionally in three fundamen-
tal ways, each of which should logically have led the United States and NATO to adopt
a more robust and considered strategy in the Kosovo war. First, the 1995 NATO air
campaign was linked to a major ground effort by Croatian and Bosnian forces coming
in from the north and west and by some 10,000 NATO troops who had been deployed
weeks prior to the onset of the bombing. In 1999, in contrast, the ground element was
expressly ruled out at the highest levels. Second, the objective of Deliberate Force was
limited (ending the siege of Sarajevo) and achievable through a phased, coercive
bombing campaign, whereas the goals of Allied Force were ambiguous (including
forcing Milosevic back to the bargaining table) and more difficult to achieve through
air power alone. Finally, even before the onset of the 1995 bombing, Milosevic had
told U.S. negotiators that he was interested in forging a deal to end the war in Bosnia
on terms acceptable to the international community. That was anything but the case
on the eve of Allied Force. I thank Ivo Daalder of the Brookings Institution for calling
my attention to these differences.
14Michael Dobbs, “‘Europe’s Last Dictator’ Digs In,” Washington Post, April 26, 1999.
Lapses in Strategy and Implementation 185
insisted that Operation Allied Force “could not have been conducted
without the NATO alliance and without the infrastructure, transit
and basing access, host-nation force contributions, and most impor-
tant, political and diplomatic support provided by the allies and
other members of the coalition.”15 Yet the conduct of the air war as
an allied effort, however unavoidable it may have been, came at the
cost of a flawed strategy that was further hobbled by the manifold
inefficiencies that were part and parcel of conducting combat opera-
tions by committee.
______________
15Secretary of Defense William S. Cohen and General Henry H. Shelton, “Joint State-
ment on the Kosovo After-Action Review,” testimony before the Senate Armed Ser-
vices Committee, Washington, D.C., October 14, 1999.
16Thomas W. Lippman and Bradley Graham, “Yugoslavs Fire on U.S. Troops; 3 Miss-
ing,” Washington Post, April 1, 1999.
17Bob Deans, “Pentagon Mum About Air Mission,” European Stars and Stripes, April
27, 1999.
186 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Daily target production began at the U.S. Joint Analysis Center at RAF
Molesworth, England, where analysts collated and transmitted the
latest all-source intelligence, including overhead imagery from satel-
lites and from Air Force Predator, Navy Pioneer, and Army Hunter
UAVs. Because the United States commanded the largest number of
intelligence assets both in the theater and worldwide by a substantial
margin, it proposed most of the targets eventually hit, although other
allies made target nominations as well.19 With the requisite infor-
mation in hand, target planners at SHAPE and USEUCOM would
then begin assembling target folders, conducting assessments of a
proposed target’s military worth, and taking careful looks at the like-
lihood of collateral damage. In addition, lawyers would vet each
proposed target for military significance and for conformity to the
law of armed conflict as reflected in the Geneva Conventions.
Once ready for review and forwarding up the chain of command for
approval, these target nominations would then go to the Joint Target
______________
18 Michael R. Gordon, “Allied Air Chief Stresses Hitting Belgrade Sites,” New York
Times, May 13, 1999.
19General Wesley Clark, USA, testimony to the Senate Armed Services Committee,
Washington, D.C., July 1, 1999.
Lapses in Strategy and Implementation 187
______________
20John A. Tirpak, “The First Six Weeks,” Air Force Magazine, June 1999, pp. 27–29.
21Dana Priest, “Target Selection Was Long Process,” Washington Post, September 20,
1999.
22Comments on an earlier draft by Hq USAFE/SA, April 6, 2001. As a rule, the 19 indi-
vidual allies did not deliberate over every new target added to the list. True enough,
the NAC—that is, all 19 members, from the United States to Luxembourg—had to
agree to move from one phase in the air war to the next. On January 30, 1999, for ex-
ample, the NAC authorized NATO’s secretary general to commence Phase I (attacking
the IADS and some command and control targets) whenever diplomatic efforts had
been deemed exhausted (as it turned out, on March 24, when Solana finally ordered
Clark to begin the bombing). The NAC also approved moving to Phase II on March 27,
thereby allowing NATO to strike against military targets north of the 44th parallel. Al-
though it never approved Phase III, which entailed strikes against military targets
throughout the former Yugoslavia, the NAC gave de facto approval to entering this
phase on March 30. From that point on, aside from Britain, France, and the United
States, no NATO country ever reviewed, let alone approved or vetoed, any individual
weapon aim point. France insisted on reviewing targets in Montenegro; Britain,
France, and the United States all demanded the right to review any target that had
high political significance or was located in or near civilian areas where the risks of
collateral damage were significant. But the remainder of the allies only got to vote on
proposed new target categories. Moreover, targets struck by U.S. aircraft operating
outside NATO but within USEUCOM were not subject to outside review unless they
met these two criteria.
188 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
23This problem will only get worse as the low-observable F-22 and Joint Strike Fighter
begin coming on line in significant numbers toward the end of this decade. Should
the United States intend to use these third-generation stealth aircraft in a coalition
context, as seems to be most likely, a dual ATO arrangement of the type used in Allied
Force will not work. New standardized tactics, techniques, and procedures will need
to be perfected and employed regularly in routine allied and combined peacetime
training. I am grateful to my RAND colleagues James Schneider, Myron Hura, and
Gary McLeod for this important insight.
24John A. Tirpak, “Short’s View of the Air Campaign,” Air Force Magazine, September
1999.
Lapses in Strategy and Implementation 189
RAND MR1365-7.1
NATO U.S.
Combined Forces Joint Forces
Air Component ATO Process Air Component
Commander’s
Intent
Joint
5th Allied Tactical Air Operations Center
Air Force
Combat Plans
Strategy and
Division
Measures of Merit
Combined Strategic
Air Operations Center Planning
Combat Combat
Operations ATO Production Operations
Division and Distribution Division
SOURCE: Hq USAFE/SA.
Once the initial hope that Milosevic would fold within a few days af-
ter the bombing commenced was proven groundless, NATO was
forced into a scramble to develop an alternative strategy. The im-
mediate result was an internecine battle between Clark and his Air
Force subordinate over where the air attacks should be directed.
Short had naturally chafed from the very beginning at the slowness of
Operation Allied Force to gather momentum—three successive
______________
25For the pertinent details of that controversy, see Benjamin S. Lambeth, The Trans-
formation of American Air Power, Ithaca, New York, Cornell University Press, 2000, pp.
130–138.
26General Wesley K. Clark, Waging Modern War: Bosnia, Kosovo, and the Future of
Combat, New York, Public Affairs, 2001, pp. 241, 243–244. As for charges of “alleged
micromanagement,” Clark said only that he many times “found [himself] working
further down into the details than [he] would have preferred, in an effort to generate
the attack effectiveness against the ground forces that [he] knew we needed.” Ibid., p.
245. Short’s countervailing take on all this is presented in candid detail in Lieutenant
General Michael C. Short, USAF (Ret.), “An Airman’s Lessons from Kosovo,” in John
Andreas Olsen, ed., From Maneuver Warfare to Kosovo, Trondheim, Norway, Royal
Norwegian Air Force Academy, 2001, pp. 257–288.
Lapses in Strategy and Implementation 191
nights had been required just to get through the 51 targets that had
been approved up to that point, most of them air defense-related
and only a few located anywhere in or near Belgrade.27 In light of the
absence of an allied ground threat to flush out Serbia’s dispersed and
hidden forces in Kosovo, Short insisted that a more effective use of
allied air power would be to pay little heed to those forces and to
concentrate instead on infrastructure targets in and near downtown
Belgrade and other cities, including key electrical power plants and
government ministries.
______________
27Of these initial approved targets, 35 were IADS-related, seven entailed VJ and MUP
facilities, seven involved command and control nodes, and two were industrial.
Comments on an earlier draft by Hq USAFE/HO, May 10, 2001.
192 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
28William M. Arkin, “How Sausage Is Made,” Washington Post, July 17, 2000. Clark
himself later affirmed in a backhanded way that he regarded General Short more as a
subordinate to be managed than as a source of trusted counsel on air employment
matters, and that he looked instead to Short’s immediate Air Force superior, General
Jumper, for the latter: “My real window on the operation was going to be provided by
the senior American airman in Europe, John Jumper. Although he wasn’t in the NATO
chain of command for this operation, as the senior American airman he was my ad-
viser and had all the technology and communications to keep a real-time read on the
operations. As Mike Short’s commander in the American chain of command, he also
had a certain amount of influence in an advisory capacity.” Clark, Waging Modern
War, p. 195.
29Tirpak, “Short’s View of the Air Campaign.”
30Interview with Lieutenant General Michael Short, USAF, PBS Frontline, “War in Eu-
rope,” February 22, 2000. Short also later indicated his belief that the use of VTCs
“improperly allowed senior leadership to reach down to levels they did not need to be
involved in.”
31In one reported exchange during a daily video teleconference, Clark insisted that
NATO air power remain committed against enemy fielded forces in Kosovo, and Short
countered that such missions were a waste of assets and should be supplanted by
missions against downtown Belgrade. Noting that U.S. aircraft were about to attack
the Serbian special police headquarters in Belgrade, Short said: “This is the jewel in
Lapses in Strategy and Implementation 193
That said, Clark had the option all along of leaving the day-to-day
operational responsibilities of planning and implementing the air
effort to his JTF commander, Admiral Ellis, as the principal subordi-
nate warfighting CINC. That is what U.S. Army General George Joul-
wan had done as SACEUR in 1995 with Admiral Leighton Smith dur-
ing Operation Deliberate Force, so he could devote his full time,
attention, and energy to his paramount duties as a diplomat in uni-
form. Instead, Clark elected not only to shoulder his diplomatic bur-
dens as NATO’s supreme commander, but also to conduct the air
war himself from Brussels, in the process bypassing not only Admiral
Ellis but also his air component commander, General Short, in mak-
ing air apportionment decisions. Whereas General H. Norman
Schwarzkopf, on the eve of Desert Storm, had become wholly per-
suaded by his trusted JFACC, then–Lieutenant General Horner, of the
merits of the chosen air campaign strategy, Clark would not be
moved by Short from his less trusting insistence that the VJ’s 3rd
_____________________________________________________________
the crown.” To which Clark replied: “To me, the jewel in the crown is when those B-
52s rumble across Kosovo.” Short: “You and I have known for weeks that we have
different jewelers.” Clark: “My jeweler outranks yours.” Dana Priest, “Tension Grew
with Divide in Strategy,” Washington Post, September 21, 1999.
32Richard K. Betts, “Compromised Command: Inside NATO’s First War,” Foreign Af-
fairs, July/August, 2001, p. 126.
194 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
33During a 10th-anniversary retrospective featuring Schwarzkopf’s principal deputies
in Desert Storm, Horner was emphatic on the crucial importance of the ability of those
key deputies to work together harmoniously in producing the war’s successful
outcome: “The one thing you need to understand if you’re going to understand Desert
Storm is that the relationship among the four people at this table—[Admiral Stanley]
Arthur, [General Walter] Boomer, [Lieutenant General John] Yeosock, and me—was
highly unusual. Such a relationship probably has never existed before, and it probably
won’t exist in the future. The trust and respect we had for one another was
unbelievable. This was a function of personality as much as a desire to get the job
done. Unless you understand our relationships, then you really won’t understand
what went on in Desert Storm, all the good and bad—and there was plenty of each.”
“Ten Years After,” Proceedings, January 2001, p. 65.
34R. W. Apple, Jr., “With Decision to Attack, a New Set of U.S. Goals,” New York Times,
March 25, 1999.
35“Stumbling into War,” The Economist, March 27, 1999, p. 17.
Lapses in Strategy and Implementation 195
______________
36James Gerstenzang and Elizabeth Shogren, “Serb TV Airs Footage of 3 Captured U.S.
Soldiers,” Los Angeles Times, April 1, 1999.
37Jonathan Foreman, “The Casualty Myth,” National Review, May 3, 1999, p. 40.
38Short, interview on PBS Frontline.
39Ibid.
196 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Short later declined even to give Allied Force the courtesy of calling it
a “campaign,” saying that it was not an operation aimed at achieving
clear-cut strategy goals with dispatch, but rather something more in
the nature of “random bombing of military targets.”40 It was one
thing, Short said, to go after enemy tanks and APCs in the Iraqi desert
the way the coalition did with such success in Desert Storm before
the ground offensive began. In that instance, everything behind the
forward edge of the battle area was enemy territory, where one could
attack targets at will without concern for collateral damage or the
potential for killing refugees. In the contrasting case of Kosovo, he
said, “we felt that the risk was enormous, and we felt that we were
going to spend a lot of assets to get minimum return. It was going to
take a lot of sorties to kill a tank, and there was enormous risk of hit-
ting the wrong target because we knew refugees would be moving
around in this ethnic cleansing environment.” Short’s preference
was to “go after the head of the snake,” as he put it. In an illustration
of what he meant, he suggested that ten combat sorties against Bel-
grade would all hit their targets and achieve a desired effect, whereas
“if I send those same ten sorties into Kosovo, perhaps we’ll find a
tank, perhaps not, [and] if we don’t, we send the ten sorties to what
in my business we call a ‘dump target,’ which is a suspected assem-
bly area or a barracks from which the enemy has fled two weeks ago,
and we’ll blow up empty buildings. So the bombs will hit something
but the impact on ethnic cleansing is zero.”
For their part, NATO’s civilian leaders could not even bring them-
selves to face the fact that they were engaged, to all intents and pur-
poses, in an ongoing war. Three weeks into Allied Force, Secretary
Cohen declared before the Senate Armed Services Committee:
“We’re certainly engaged in hostilities. We’re engaged in combat.
Whether that measures up to, quote, a classic definition of war, I’m
not prepared to say.”41 Such diffidence on the administration’s part
was ostensibly intended to reflect due executive-branch obeisance to
______________
40Ibid.
41“Verbatim Special: The Balkan War,” Air Force Magazine, June 1999, p. 50.
Lapses in Strategy and Implementation 197
______________
42Bill Gertz and Rowan Scarborough, “No War,” Washington Times, April 16, 1999.
43“Verbatim Special: The Balkan War,” p. 51.
44Charles Babington, “Clinton Sticks with Strikes as Poll Shows 51 Percent in U.S. Ap-
prove,” Washington Post, March 30, 1999.
45Richard Benedetto, “Support Not as High as for Other Strikes,” USA Today, April 2,
1999.
198 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
just cause,” he pointed out, “the public applauds the use of force.
When it loses—worse still, when America is defeated or runs away (as
in Somalia or Vietnam)—the public reasonably says the use of the
military was a mistake.” Citing the precedent of Desert Storm, he re-
called how during the days immediately preceding the outbreak of
hostilities, no poll found a majority of Americans in favor of prompt
military action. Yet immediately after the air campaign had begun
and was deemed to have gotten off to a good start, surveys found that
between 68 and 84 percent of those polled approved. Similarly, up to
the day before the Desert Storm ground push commenced, a typical
poll taken by the New York Times and CBS found that the public pre-
ferred a continuation of the air war by 79 percent, with only 11 per-
cent favoring the start of ground operations. A few days after the
ground push began, however, a full 75 percent of those polled be-
lieved it had been “right to start the ground war,” as opposed to only
19 percent who opposed it.46
______________
46Humphrey Taylor, “Win in Kosovo and the Public Will Approve,” Wall Street Jour-
nal, June 3, 1999.
47James Cox, “Poll: Mission Isn’t Seen as U.S. Victory,” USA Today, June 15, 1999.
Lapses in Strategy and Implementation 199
______________
48Ignatieff, Virtual War, p. 99.
200 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
49Elaine M. Grossman, “U.S. Military Debates Link Between Kosovo Air War, Stated
Objectives,” Inside the Pentagon, April 20, 2000, p. 7. According to one Allied Force
participant, Clark would press his operators down the line to propose target candi-
dates. They would reply, “Give us the targets and we will take them out,” to which
Clark countered: “You don’t get it. You develop the targets.” Quoted in Ignatieff, Vir-
tual War, p. 99. Clark himself later justified 2,000 as “a large round number, large
enough to get us past the daily struggle over the number of targets approved for that
day.” Clark, Waging Modern War, p. 250.
50William M. Arkin, “Smart Bombs, Dumb Targeting?” Bulletin of the Atomic Scien-
tists, May/June 2000.
51Ibid.
Lapses in Strategy and Implementation 201
NATO. 52 In the words of another officer, “nobody ever said, ‘no fool-
ing, what we want to accomplish in this country is X.’” As a result,
NATO started “throwing bombs around, hoping that objectives
would materialize.” Said still another, “the targets we selected—
because we had no objectives—were based on nothing other than
that they had been approved. So we slung lead on targets [but] we
couldn’t say, ‘the objectives are X, so we blew up Y.’”53
______________
52Personal communication to the author, August 23, 1999.
53Grossman, “U.S. Military Debates Link Between Kosovo Air War, Stated Objectives,”
p. 8.
54Comments at an Air Force Association Eaker Institute colloquy, “Operation Allied
Force: Strategy, Execution, Implications,” held at the Ronald Reagan International
Trade Center, Washington, D.C., August 16, 1999. When asked about effects-based
targeting applications in Allied Force, the former commander of the Joint Warfare
Analysis Center, which provides senior warfighters with the principal analytical
202 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
After the dust of Operation Allied Force had settled, the since-retired
commandant of the U.S. Marine Corps, General Charles Krulak,
_____________________________________________________________
support for such targeting, remarked, “the campaign was more like random acts of
violence than true effects-based targeting. The legal restrictions and political
constraints in the target approval process were inexplicably given as excuses not to do
effects-based targeting. Achieving the desired effects while minimizing the undesired
effects, particularly under the restrictions and constraints that were placed on
SACEUR, is precisely why effects-based targeting should have been applied. Anything
else is just high-tech vandalism.” Conversation with Captain C. J. Heatley, USN (Ret.),
Arlington, Virginia, June 21, 2000.
55John A. Tirpak, “Kosovo Retrospective,” Air Force Magazine, April 2000, p. 31.
Lapses in Strategy and Implementation 203
commented from firsthand involvement that “we did not have a real
strategy.”56 Likewise, General Short remarked, in what was surely an
understatement for him, that the bombing effort had produced its
objectives “to some extent by happenstance rather than by de-
sign.”57 There were later intimations that a hidden agenda of both
the Clinton administration and General Clark had been not just a re-
versal of the ethnic cleansing in Kosovo, but nothing less than the
removal of Milosevic from power and the democratization of
Yugoslavia. On that point, one NATO official later described Clark as
having said, “you must understand that the objective is to take
Yugoslavia away from Mr. Milosevic, so we can democratize it and
modernize it. That’s our objective.”58 But it was never communi-
cated to subordinate staffs or made a declared goal of Allied Force.59
Given the unseemly rush for targets that ensued at SHAPE and else-
where for more than a month after NATO’s initial assumptions
proved groundless, it seemed more than a bit disingenuous for ad-
ministration officials to have claimed afterward that although they
had “hoped” that military action would end the Serb abuses in
Kosovo quickly, “we knew that it was equally possible that it would
not and that a sustained campaign might be necessary to stop the
killing and reverse the expulsions” and that “we were prepared to do
what it took to win.”60 In what bore every hallmark of a post-hoc at-
tempt at historical revisionism, one official professed that “people in
Washington” knew that there would be a need to attack infrastruc-
ture targets once it became clear that a three- to four-day bombing
effort would not compel Milosevic to settle, but because the allies,
especially the French, were “not on board” initially, NATO could not
______________
56Grossman, “U.S. Military Debates Link Between Kosovo Air War, Stated Objectives,”
p. 6.
57Tirpak, “Kosovo Retrospective,” p. 33.
58Quoted in Grossman, “U.S. Military Debates Link Between Kosovo Air War, Stated
Objectives,” p. 7.
59General Krulak later remarked that even had it been an unstated goal, it was a “non-
starter,” because it would never have gained the backing of NATO.
60James B. Steinberg, “A Perfect Polemic: Blind to Reality on Kosovo,” Foreign Affairs,
November/December 1999, p. 131.
204 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
start attacking Phase III targets until it had consensus about the
bombing.61 Two critics of administration policy countered convinc-
ingly that such claims by the administration that it had been pre-
pared all along for the possible need for a prolonged air campaign
were flatly belied by “the hasty improvisation that marked the
bombing effort.” 62 True enough, General Clark was said on strong
authority never to have suggested that just a few days of bombing
would suffice to do the job, even though he did limit his planners to a
short-duration operations plan out of a conviction that the alliance’s
political leaders would not sit still for anything longer.63 But the pre-
sumptions of both NATO and the most senior officials of the Clinton
administration were well reflected in U.S. interagency reports in Jan-
uary and February 1999, which argued confidently that “after enough
of a defense to sustain his honor and assuage his backers, [Milosevic]
will quickly sue for peace.”64
______________
61This official further claimed that Clinton had never intended to take the ground op-
tion off the table but “downplayed” it at first on the grounds that any public mention
of it could have prompted a bruising debate in Congress and premature pressures to
invoke the War Powers Act. He added that by April, the administration felt compelled
to change that perception when it had become clear that important audiences had
concluded that the president had flatly ruled out any ground option. That attempt to
shift perceptions, he said, included asking Solana to initiate a review of the forces that
would be required and encouraging Clark to accelerate planning for a ground inva-
sion, making no effort to keep this quiet. The official admitted that there was no way a
ground invasion could have been imminent when Milosevic capitulated on June 3, but
that any decision to proceed with an invasion most definitely would have had to be
made by mid-June so that the logistical provisions needed to support a ground offen-
sive could be completed before the onset of winter. Interview by RAND staff, Wash-
ington, D.C., July 11, 2000.
62Christopher Layne and Benjamin Schwarz, “Kosovo II: For the Record,” The Na-
tional Interest, Fall 1999, p. 12.
63Conversation by RAND staff with Lieutenant General Ronald Keys, USAF, director of
operations, U.S. European Command, Stuttgart, Germany, March 8, 2000. Clark him-
self was clear on this point in his subsequently published memoirs. Although he ac-
knowledged that “there was a spirit of hope at the political levels [going into the
bombing] that Milosevic might recognize that NATO was actually going to follow
through with its threat and then quickly concede in order to cut his losses,” he, for his
own part, suspected all along that it was “going to be a long campaign.” Clark, Waging
Modern War, pp. 177, 201.
64Elaine Sciolino and Ethan Bronner, “How a President, Distracted by Scandal, En-
tered Balkan War,” New York Times, April 18, 1999.
Lapses in Strategy and Implementation 205
A senior NATO official commented that “NATO got in way over its
head, stumbled through, didn’t know how to get out, [and] was
______________
65John M. Broder, “How to Lay Doubt Aside and Put the Best Face on a Bad Week in
the Balkans,” New York Times, April 1, 1999.
66David A. Fulghum, “Lessons Learned May Be Flawed,” Aviation Week and Space
Technology, June 14, 1999, p. 205. Even deeper than the problem of slow target ap-
proval, however, was the problem of positive target identification, given the excep-
tional stringency of the prevailing rules of engagement. For example, Joint STARS
could not distinguish a column of refugees from a column of military vehicles loaded
with enemy troops, a performance shortfall far more difficult to fix than streamlining
the approval process. I am grateful to my colleague Bruce Pirnie for pointing this out.
206 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
scared to death by what was happening.” This official added that the
entire bombing effort had been a “searing experience” that had “left
a bitter taste of tilting within governments, between governments,
between NATO headquarters in Brussels and the military headquar-
ters at Mons.”67 Reflecting the consensus arrived at by many senior
U.S. military officers, both active and retired, Admiral Leighton
Smith concluded that “the lesson we’ve learned is that coalitions
aren’t good ways to fight wars” and that, at a minimum, the political
process in NATO needed to be streamlined so that the collective
could use force in a way that made greatest military sense.68
Even before that event, the Pentagon had admitted the discovery of
operational security problems, as well as its suspicions that the Serbs
had gained access to at least parts of the ATO, thereby enabling them
to reposition mobile SAMs in anticipation of planned attacks.70 Alle-
gations that France, in particular, had been kept out of the loop with
respect to some target planning because of concern that the infor-
mation would be passed on to Milosevic were tacitly confirmed in
early April by a Clinton confidant, who remarked that “there are cir-
______________
67Jane Perlez, “For Albright’s Mission, More Problems and Risk,” New York Times,
June 7, 1999.
68Bradley Graham and Dana Priest, “‘No Way to Fight a War’: Hard Lessons of Air
Power, Coalitions,” Washington Post, June 6, 1999.
69Hugo Gurdon, “U.S. Admits Milosevic Spies Are Inside NATO,” London Daily Tele-
graph, April 15, 1999.
70 Roberto Suro and Thomas E. Ricks, “Pentagon: Kosovo Air War Data Leaked,”
Washington Post, March 10, 2000.
Lapses in Strategy and Implementation 207
______________
71 Hugo Gurdon, “France Kept in Dark by Allies,” London Daily Telegraph, April 9,
1999.
72“NATO Chief: Targeting Goals Leaked to Yugoslavia,” Pacific Stars and Stripes, Au-
gust 13, 1999.
73Tom Raum, “Cohen: NATO Process Prolonged Air Strikes,” Philadelphia Inquirer,
July 21, 1999.
74Other examples of such self-inflicted wounds, in Ellis’s view, were excessively high
standards for limiting collateral damage, NATO’s self-suspension of the use of cluster
munitions, the aversion to casualties and ground combat, and the reactive as opposed
to proactive public affairs posture, all of which slowed allied response time and re-
duced allied control over the air war’s operational tempo.
75Quoted in Lieutenant Colonel L. T. Wight, USAF, “What a Tangled Web We Wove:
An After-Action Assessment of Operation Allied Force’s Command and Control Struc-
ture and Processes,” unpublished paper, no date, p. 1. Colonel Wight was a member
of the C-5 Strategy Cell at the CAOC.
208 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
dnaNorth
mmoC Atlantic
lanoitaN National
dnammoCommand
C lanoitaN
seCouncil
itirohtuA Authorities
seitirohtuA
NATO
.S.U U.S.
.S.U
SOURCE: Hq USAFE/SA.
______________
76Although the USAF’s still-embryonic Air Expeditionary Forces were not available for
participation in Operation Allied Force, the AEF concept was exercised at Aviano when
the reinforced 31st Fighter Wing was designated a provisional air expeditionary wing
for the air war’s duration.
77Complicating matters even further was the added confusion created by having Task
Force Hawk and JTF Shining Hope functioning as separate command entities within
the joint operating area.
210 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
78Wight, “What a Tangled Web We Wove,” p. 7. As a case in point, the Joint Chiefs,
despite their formal status as advisers to the NCA, issued directives as though they
were part of the warfighting chain of command, for instance, ruling out the use of
CBUs by U.S. forces and placing certain targets on “JCS withhold.” Likewise, JTF No-
ble Anvil was often placed in a position of providing direction and guidance to NATO
operational units, even though it nominally exercised operational and tactical control
over U.S. assets only.
Lapses in Strategy and Implementation 211
Jumper had “a major say in how the huge USAF contribution was
used.”79 Moreover, unlike General Horner in Desert Storm, who an-
swered directly to the theater CINC, General Schwarzkopf, Short re-
ported not to Clark but rather to Admiral Ellis, who in turn reported
to Clark—a situation which, Short cautiously said, “colors the equa-
tion a bit in terms of my latitude, if you will, in this air campaign.”80
Considering all this confusion and more, concluded an informed and
expert observer, operational effectiveness in Allied Force was prob-
ably achieved “in spite of the . . . command structures and processes
rather than as a direct result of them.”81
Even when more fully developed, the BDA process left much to be
desired. It was well enough equipped, calling as required on national
and theater offboard sensors such as satellites and the U-2, tactical
sensors such as Predator and Hunter UAVs, and onboard sensors
such as the LANTIRN targeting pod carried by the F-14D, the F-15E,
and the Block 40 F-16CG. Inputs from these information sources
would be forwarded to the JAC at RAF Molesworth, which wielded
chief BDA authority, and other BDA-related entities such as the
CAOC, national agencies, the SACEUR staff, JTF Noble Anvil, and the
JFACC apparatus. However, inputs from national and theater assets
could take days to register an impact because of frequent weather
complications and higher-priority taskings. Moreover, because BDA
______________
79Air Commodore A. G. B. Vallance, RAF, chief of staff, NATO Reaction Forces (Air)
Staff, Kalkar, Germany, “After Kosovo: Implications of Operation Allied Force for Air
Power Development,” unpublished paper, p. 3. Although Clark did, by numerous
eyewitness accounts, sometimes treat Jumper as though he were the air component
commander by virtue of his seniority to Short, Jumper never usurped his superior
rank, never insisted that Short follow his suggestions, and frequently lent a helpful
hand by quietly adjudicating the more prickly VTC sessions to good effect when Clark
and Short got into their differences over targeting strategy and target priorities.
80Michael R. Gordon, “Allied Air Chief Stresses Hitting Belgrade Sites.”
81Wight, “What a Tangled Web We Wove,” p. 1.
212 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
82 Briefing to the author by Brigadier General Daniel J. Darnell, commander, 31st
Fighter Wing, Aviano Air Base, Italy, June 13, 2000.
83Wight, “What a Tangled Web We Wove,” p. 9.
Lapses in Strategy and Implementation 213
______________
84The problem was not just the absence of a land component per se, but that no com-
ponent whatsoever undertook the task of IPB until far too late in the operation. What
is required are clearer stipulations regarding whose responsibility it is to conduct IPB,
as well as new approaches and processes for doing so. At present, only the land
component is resourced and prepared to meet that responsibility. Comments on an
earlier draft by Hq USAF/XOXS, July 11, 2001.
85Rowan Scarborough, “Kosovo Target Data Stalled in Transit,” Washington Times,
July 28, 1999.
214 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Over time, the CAOC went from badly understaffed to packed with a
surfeit of personnel as a result of the rampant inefficiencies of the
target planning and apportionment process. On one occasion, there
were as many as 1,400 people in the small and cramped facility, pro-
ducing a staffing level that bordered on gridlock.87 Some aug-
mentees from other USAF commands brought only limited experi-
ence with high-intensity operations, further hampering the CAOC’s
operational effectiveness. In a representative example of the need-
less inefficiencies that ensued, a PACAF colonel, say, serving as se-
nior duty officer, would overrule something decided at a lower level
with a “we don’t do that in PACAF,” only to have a lieutenant colonel
on the permanent CAOC staff reply, uneasily, that that was the way it
was done in Allied Force, for good reason.88 In general, the abnor-
mally large number of senior officers (lieutenant colonels and
colonels) populating the CAOC limited the effectiveness of the often
more expert junior officers in shaping key decisions. As a rule, the
CAOC and General Short mainly performed battle management and
support functions rather than operating as a master planning center
and high-level command and control entity along the lines of the Air
Operations Center and General Horner in Desert Storm.
Yet for all its eventually ramped-up staffing and improved organiza-
tion, according to its director at the time, the CAOC remained “target
______________
86As for information operations, one Allied Force participant commented that “due to
the involvement of a few compartmented programs, the entire planning effort was
classified at an unnecessarily high level, unreleasable to all but a very few U.S. plan-
ners. Unfortunately, implementing the overall plan was critical to the success of the
operation, but because of the excessive classification, those charged with implement-
ing it could not be told of the plan until it was too late.”
87The CAOC’s normal peacetime manning was around 250 assigned personnel. It had
a reinforced staff of 375 on March 24, the night the air war began, which was finally
ramped up to more than 1,400 as Allied Force peaked at more than 900 sorties a day.
88Conversation with Major General P. J. M. Godderij, deputy commander in chief,
Royal Netherlands Air Force, Scheveningen, the Netherlands, June 7, 2000.
Lapses in Strategy and Implementation 215
As for the flexible targeting effort against VJ forces in the KEZ, the
CAOC at first lacked any on-hand Army expertise to help develop the
ground order of battle. With no land component in place, the Army’s
TPQ-36 and TPQ-37 counterbattery radars in Albania required a di-
rect feed to the CAOC, yet information from them was not provided
until the very end because Army doctrine had planned for those sys-
tems to be used in a different manner and the CAOC was not config-
ured to take advantage of them. Worse yet, TF Hawk and its parent
command, the U.S. Army in Europe (USAREUR), consciously elected
not to provide processed intelligence data to the JFACC and JTF
Noble Anvil until circumstances and senior-official intervention
occurred later.90 Eventually, Clark sent a 10-man Army team to the
CAOC to provide such assistance, which aided considerably in the
flexible targeting effort. By mid-May, TF Hawk finally began sending
the CAOC useful real-time targeting information collected by its
counterbattery radars, and a battlefield coordination element staffed
with TF Hawk representatives was established in the CAOC to
provide additional ground intelligence and operator input into the
flexible targeting cell concerned with dispersed and hidden enemy
forces in the KEZ.
______________
89Brigadier General Randy Gelwix, USAF, “Oral Histories Accomplished in Conjunc-
tion with Operation Allied Force/Noble Anvil.”
90During an after-action presentation by the USAREUR battlefield coordination ele-
ment, Hq USAFE’s AWOS study team learned that JTF Noble Anvil had prepared a
memorandum of agreement for USAREUR coordination expressly stipulating that TF
Hawk would provide the CAOC with processed intelligence data from the TPQ-36 and
TPQ-37 counterbattery radars. In the ensuing coordination process, the USAREUR in-
telligence directorate reportedly excised pertinent language from the text. Comments
on an earlier draft by Hq USAF/XOXS, July 11, 2001.
216 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
91Amplifying on this point a year after the air war ended, Ellis further remarked that
because air power had been the only force element actively used in Allied Force, the
JFACC naturally had a heavy air emphasis. Yet, he added, the planning and execution
system badly needed land and maritime component commanders deep in the loop as
well, so they could explain to the JFACC, as authoritative equals, what their services
were able to bring to the planning table. Noting how the “J” in JFACC was all too often
silent, Ellis recalled that the contributions of other services were not invariably made
the best use of. For example, he said, the EA-6B, TLAM, and F-14 TARPS all brought
good capabilities to the fight and the JFACC needed to know about those capabilities
directly from their most senior operators. TARPS, in particular, offered excellent po-
tential value, but the Air Force, now out of the manned tactical reconnaissance busi-
ness, sometimes gave the impression of believing that if the information did not come
from space, it did not have an obvious use. Ellis’s overall point was that the services
have not yet become sufficiently joint-minded at the operational and tactical levels, let
alone the strategic level. Interview with Admiral Ellis, May 30, 2000.
92 Conversation with Vice Admiral Daniel J. Murphy, USN, 6th Fleet commander,
aboard the USS LaSalle, Gaeta, Italy, June 8, 2000.
Lapses in Strategy and Implementation 217
with VTC sessions taking place daily at the most senior level because
of the wide geographic spread of the key players. Sometimes as
many as three or four VTCs were conducted in one day among the
most senior principals. Admiral Ellis later characterized them as a
powerful tool if properly used, owing to their ability to shorten deci-
sion cycle times dramatically, to communicate a commander’s intent
clearly and unambiguously, and to obviate any requirement for the
leading commanders to be collocated. But he cited the propensity of
VTCs to be voracious consumers of leadership and staff working
hours (often involving time wasted composing flashy but unneces-
sary—and even at times counterproductive—briefing graphics) and
poor substitutes for rigorous mission planning and written orders.
Decisions made in the VTC were all too readily prone to misinterpre-
tation as key guidance was successively handed down to lower staff
levels.93
______________
93Briefing by Admiral James O. Ellis, USN, commander in chief, U.S. Naval Forces,
Europe, and commander, Allied Forces Southern Europe and Joint Task Force Noble
Anvil, “The View from the Top,” 1999.
94Wight, “What a Tangled Web We Wove,” p. 10.
95Ibid., p. 11.
218 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
After the war ended, criticism of the VTC approach by many senior
officers was quite vocal. In a characteristic observation, the UK
Ministry of Defense’s director of operations in Allied Force, RAF Air
Marshal Sir John Day, remarked that for all its admitted efficiencies
when its use was properly disciplined, the VTC mechanism was
highly conducive to “ad-hocracy” of all sorts, sometimes resulting in
a lack of clarity regarding important matters of both planning and
execution. For example, he observed that because of the federated
nature of the operation’s planning and the extensive use of VTCs in-
volving a large number of U.S. and NATO headquarters, many agen-
cies had full knowledge of the planning details. That generated initial
confusion among the UK participants as to who precisely was
running the air war, since, until it was confirmed (as suspected) that
it was indeed General Short, they could obtain the same information
from any headquarters that was involved in the VTC. Consistent with
others who reflected on the many negatives of VTCs with the benefit
of hindsight, Air Marshal Day suggested that participation in high-
level VTCs should henceforth be limited exclusively to those directly
in the chain of command and that the commander in chief should
devote careful thought beforehand to the following: (1) the appro-
priate participants and viewers; (2) a prior agenda, so that essential
participants would not hesitate to raise an item out of fear that an
item might already be on the CINC’s checklist; (3) diligent minute-
taking; and (4) a summary of command decisions taken, so that the
commander’s intent would always be unambiguous.96
______________
96Interview with Air Marshal Sir John Day, RAF Innsworth, United Kingdom, July 26,
2000. Rather more bluntly, retired USAF General Chuck Horner, the JFACC during
Desert Storm, commented that had he been SACEUR during Allied Force, he would
have shot every TV monitor in sight. The biggest problem with VTCs, Horner said, is
that one does not know who is present and listening, even as a videotaped record of
the proceeding is being made. That, he added, inclines participants to pull their
punches and speak “for the record,” rather than to speak their mind in a manner that
only privacy can ensure. Conversation with General Horner at Farnborough, United
Kingdom, July 27, 2000.
Chapter Eight
NATO’S AIR WAR IN PERSPECTIVE
Operation Allied Force was the most intense and sustained military
operation to have been conducted in Europe since the end of World
War II. It represented the first extended use of military force by
NATO, as well as the first major combat operation conducted for
humanitarian objectives against a state committing atrocities within
its own borders. It was the longest U.S. combat operation to have
taken place since the war in Vietnam, which ended in 1975. At a
price tag of more than $3 billion all told, it was also a notably expen-
sive one.1 Yet in part precisely because of that investment, it turned
out to have been an unprecedented exercise in the discriminate use
of force on a large scale. Although there were some unfortunate and
highly publicized cases in which innocent civilians were tragically
killed, Secretary of Defense William Cohen was on point when he
characterized Allied Force afterward as “the most precise application
of air power in history.” 2 In all, out of some 28,000 high-explosive
munitions expended altogether over the air war’s 78-day course, no
more than 500 noncombatants in Serbia and Kosovo died as a direct
result of errant air attacks, a new low in American wartime experi-
ence when compared to both Vietnam and Desert Storm.3
______________
1Lisa Hoffman, “U.S. Taxpayers Faced with Mounting Kosovo War Costs,” Washington
Times, June 10, 1999.
2Bradley Graham, “Air Power ‘Effective, Successful,’ Cohen Says,” Washington Post,
June 11, 1999.
3That was the final assessment of an unofficial post–Allied Force bomb damage survey
conducted in Serbia, Kosovo, and Montenegro by a team of inspectors representing
Human Rights Watch. A U.S. Air Force analyst who was later briefed on the study
commented that Human Rights Watch had “the best on-the-ground data of anyone in
219
220 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
After Allied Force ended, air power’s detractors lost no time in seek-
ing to deprecate NATO’s achievement. In a representative case in
point, retired U.S. Army Lieutenant General William Odom charged
that “this war didn’t do anything to vindicate air power. It didn’t
stop the ethnic cleansing, and it didn’t remove Milosevic”—as
though those were ever the expected goals of NATO’s air power em-
ployment to begin with. 4 Yet because of the air war’s ultimate suc-
cess in forcing Milosevic to yield to NATO’s demands, the predomi-
nant tendency among most outside observers was to characterize it
as a watershed achievement for air power. One account called Op-
eration Allied Force “one of history’s most impressive air cam-
paigns.”5 Another suggested that if the cease-fire held, the United
States and its allies would have accomplished “what some military
experts had predicted was impossible: a victory achieved with air
power alone.”6 A Wall Street Journal article declared that Milosevic’s
capitulation had marked “one of the biggest victories ever for air
power,” finally vindicating the long-proclaimed belief of airmen that
“air power alone can win some kind of victory.”7 And the New York
_____________________________________________________________
the West.” “A New Bomb Damage Report,” Newsweek, December 20, 1999, p. 4. A
later report, however, indicated that Human Rights Watch had identified 90 separate
collateral damage incidents, in contrast to the acknowledgment by NATO and the U.S.
government of only 20 to 30. Bradley Graham, “Report Says NATO Bombing Killed 500
Civilians in Yugoslavia,” Washington Post, February 7, 2000.
4Mark Thompson, “Warfighting 101,” Time, June 14, 1999, p. 50. Regarding Odom’s
first charge, General Jumper categorically declared after the bombing effort success-
fully ended that “no airman ever promised that air power would stop the genocide
that was already ongoing by the time we were allowed to start this campaign.” Quoted
in The Air War Over Serbia: Aerospace Power in Operation Allied Force, Washington,
D.C., Hq United States Air Force, April 1, 2000, p. 19. One of the few detractors of air
power who was later moved to offer an apologia for having been wrong was military
historian John Keegan, who acknowledged a week before Milosevic finally capitulated
that he felt “rather as a creationist Christian . . . being shown his first dinosaur bone.”
John Keegan, “Modern Weapons Hit War Wisdom,” Sydney Morning Herald, June 5,
1999. Keegan, long a skeptic of air power’s avowed promise, wrote on the eve of
Milosevic’s capitulation that the looming settlement represented “a victory for air
power and air power alone.” Quoted in Elliott Abrams, “Just War. Just Means?” Na-
tional Review, June 28, 1999, p. 16.
5William Drozdiak and Anne Swardson, “Military, Diplomatic Offensives Bring About
Accord,” Washington Post, June 4, 1999.
6Paul Richter, “Air-Only Campaign Offers a False Sense of Security, Some Say,” Los
Angeles Times, June 4, 1999.
7Thomas E. Ricks and Anne Marie Squeo, “Kosovo Campaign Showcased the Effec-
tiveness of Air Power,” Wall Street Journal, June 4, 1999.
NATO’s Air War in Perspective 221
It was not just outside observers, moreover, who gave such ready
voice to that upbeat assessment. Shortly after the cease-fire, Presi-
dent Clinton himself declared that the outcome of Allied Force
“proved that a sustained air campaign, under the right conditions,
can stop an army on the ground.”10 Other administration leaders
were equally quick to congratulate air power for what it had done to
salvage a situation that looked, almost until the last moment, as
though it was headed nowhere but to a NATO ground involvement of
some sort. In their joint statement to the Senate Armed Services
Committee after the air war ended, Secretary Cohen and General
______________
8Serge Schmemann, “Now, Onward to the Next Kosovo. If There Is One,” New York
Times, June 16, 1999.
9Andrew Krepinevich, “Two Cheers for Air Power,” Wall Street Journal, June 11, 1999.
10Pat Towell, “Lawmakers Urge Armed Forces to Focus on High-Tech Future,” Con-
gressional Quarterly Weekly, June 26, 1999, p. 1564. Actually, the air effort proved no
such thing with respect to VJ forces operating in Kosovo.
222 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
With all due respect for the unmatched professionalism of those al-
lied aircrews who, against difficult odds, actually carried out the air
effort and made it succeed in the end, it is hard to accept such glow-
ing characterizations as the proper conclusions to be drawn from
Allied Force. In fact, many of them are at marked odds with the
views of those senior professionals who, one would think, would be
most familiar with air power and its limitations. Shortly before the
bombing effort began, the four U.S. service chiefs uniformly
doubted, in testimony before the Senate Armed Services Committee,
whether air strikes by themselves would succeed in compelling Milo-
sevic to yield.12 Indeed, the Air Force chief of staff, General Michael
Ryan, admitted less than a week later: “I don’t know if we can do it
without ground troops.”13 After Allied Force was over, the former
commander of NATO forces during Operation Deliberate Force, Ad-
miral Leighton Smith, remarked that the Kosovo experience should
go down as “possibly the worst way we employed our military forces
in history.” Smith added that telling the enemy beforehand what you
are not going to do is “the absolutely dumbest thing you can do.” 14
Former Air Force chief of staff General Ronald Fogleman likewise ob-
served that “just because it comes out reasonably well, at least in the
eyes of the administration, doesn’t mean it was conducted properly.
The application of air power was flawed.” Finally, the air component
commander, USAF Lieutenant General Michael Short, declared that
“as an airman, I’d have done this a whole lot differently than I was
allowed to do. We could have done this differently. We should have
done this differently.”15
______________
11Secretary of Defense William S. Cohen and General Henry H. Shelton, “Joint State-
ment on the Kosovo After-Action Review,” testimony before the Senate Armed
Services Committee, Washington, D.C., October 14, 1999.
12 Bradley Graham, “Joint Chiefs Doubted Air Strategy,” Washington Post, April 5,
1999.
13Quoted in “Verbatim Special: The Balkan War,” Air Force Magazine, June 1999, p.
47.
14“Reporters’ Notebook,” Defense Week, July 19, 1999, p. 4.
15William Drozdiak, “Allies Need Upgrade, General Says,” Washington Post, June 20,
1999.
NATO’s Air War in Perspective 223
______________
16Most others as well were caught off guard by the sudden ending of the Kosovo crisis.
See Lieutenant General Bernard E. Trainor, USMC (Ret.), “The Council on Foreign
Relations Report on the Kosovo Air Campaign: A Digest of the Roundtable on the Air
Campaign in the Balkans,” Council on Foreign Relations, New York, July 27, 2000. One
notable exception was USAF Brigadier General Daniel J. Leaf, commander of the 31st
Air Expeditionary Wing at Aviano Air Base, Italy, who confidently told his aircrews on
the eve of Milosevic’s capitulation that he could “smell an impending NATO victory in
the air” (conversation with the author in Washington, D.C., November 16, 2000).
17John F. Harris and Bradley Graham, “Clinton Is Reassessing Sufficiency of Air War,”
Washington Post, June 3, 1999.
18William M. Arkin, “Limited Warfare in Kosovo Not Working,” Seattle Times, May 22,
1999.
19General Michael E. Ryan, “Air Power Is Working in Kosovo,” Washington Post, June
4, 1999.
224 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
It hardly follows from this, of course, that air power can now “win
wars alone” or that the air-only strategy ultimately adopted by the
Clinton administration and NATO’s political leaders was the wisest
choice available to them. Yet the fact that air power prevailed on its
own despite the multiple drawbacks of a reluctant administration, a
divided Congress, an indifferent public, a potentially fractious al-
liance, a determined enemy, and, not least, the absence of a credible
NATO strategy surely testified that the air weapon has come a long
way in recent years in its relative combat leverage compared to other,
more traditional force elements. Thanks to the marked improve-
ments in precision attack and battlespace awareness, unintended
damage to civilian structures and noncombatant fatalities were kept
to a minimum, even as air power plainly demonstrated its coercive
potential.
______________
20It bears noting here that the December 1972 bombing of Hanoi was also an example
of successful coercive bombing, albeit with a very limited objective and in the context
of a much larger war that ended in defeat for the United States. For more on this, see
Wayne Thompson, To Hanoi and Back: The U.S. Air Force and North Vietnam, 1966–
1973, Washington, D.C., Smithsonian Institute Press, 2000, pp. 255–280.
21Alan Stephens, Kosovo, or the Future of War, Paper Number 77, Air Power Studies
Center, Royal Australian Air Force, Fairbairn, Australia, August 1999, p. 21.
NATO’s Air War in Perspective 225
In contrast to Desert Storm, the air war’s attempts at denial did not
bear much fruit in the end. Allied air attacks against dispersed and
hidden enemy forces were largely ineffective, in considerable part
because of the decision made by NATO’s leaders at the outset to
forgo even the threat of a ground invasion. Hence, Serb atrocities
against the Kosovar Albanians increased even as NATO air opera-
tions intensified. Yet ironically, in contrast to the coalition’s ulti-
mately unsuccessful efforts to coerce Saddam Hussein into submis-
sion, punishment did seem to work against Milosevic, disconfirming
the common adage that air power can beat up on an adversary indef-
initely but rarely can induce him to change his mind.
______________
22Michael Mandelbaum, “A Perfect Failure: NATO’s War Against Yugoslavia,” Foreign
Affairs, September/October 1999, p. 2. That charge was based on the fact that prior to
the air war’s start on March 24, 1999, only some 2,500 civilian innocents had died in
the Serb-Albanian civil war, whereas during the 11-week bombing effort, an estimated
10,000 civilians were killed by marauding bands of Serbs unleashed by Milosevic in di-
rect response to Allied Force.
226 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
23Robert Hewson, “Operation Allied Force: The First 30 Days,” World Air Power Jour-
nal, Fall 1999, p. 24.
24“Briefing by SACEUR General Wesley Clark,” Brussels, NATO Headquarters, April
13, 1999.
25 See, for example, the riposte to Mandelbaum by the Clinton administration’s
deputy national security adviser, James B. Steinberg, “A Perfect Polemic: Blind to
Reality on Kosovo,” Foreign Affairs, November/December 1999, pp. 128–133.
NATO’s Air War in Perspective 227
RAND MR1365-8.1
800
700
Total
600
500
Thousands
Albania
400
300
Macedonia
200
0
23
20
18
15
29
13
27
1
ril
ay
ne
st
ch
ril
ay
ne
ly
ly
gu
Ap
Ju
Ju
Ju
Ap
M
ar
Ju
Ju
Au
M
Reflecting on the air war experience a year later, Admiral James Ellis,
the commander of the U.S. contribution to Allied Force, observed
that during the final days leading up to March 24, it was a question
not of how the bombing effort would be conducted so much as
whether it would take place at all. Before Rambouillet, the challenge
had been to compel Milosevic to do something. Afterward, it became
to compel him to stop doing something. Ellis speculated that had the
allies known from the outset that they were signing up for a 78-day
campaign, they might easily have declined the opportunity forth-
______________
26Javier Solana, “NATO’s Success in Kosovo,” Foreign Affairs, November/December
1999, p. 114.
27Karl Mueller, “Deus ex Machina? Coercive Air Power in Bosnia and Kosovo,” un-
published paper, School of Advanced Air Power Studies, Maxwell AFB, Alabama,
November 7, 1999, p. 6.
28“Wesley Clark Looks Back,” National Journal, February 26, 2000, p. 612.
NATO’s Air War in Perspective 229
with. Unlike the ad hoc group of nations that fought Desert Storm as
a solidly united front, NATO was not a coalition of the willing but
rather a loose defensive alliance of 19 democracies. They were all
strongly inclined to march to different drummers, and all had varying
commitments to grappling—at least militarily—with humanitarian
crises in which they had no clear national security stake.29
As the bombing entered its third month without a clear end in sight,
Ellis feared that allied cohesion might collapse within three weeks
unless something of a game-changing nature occurred, such as a
drastic move by Milosevic to alter the stakes or a firm U.S. decision to
accede to a ground-invasion option. Offsetting that fear, however,
was his belief that the allies were finally beginning to recognize and
accept the need to come to terms with some thorny operational is-
sues such as granting approval to attack electrical power and other
key infrastructure targets. That took time, Ellis said, but the fact that
it finally occurred constituted a signal that the alliance was slowly
learning how to do what needed to be done.
Finally, for all the criticism that was directed against some of the less
steadfast NATO members for their rear-guard resistance and ques-
tionable loyalty while the air war was under way, even the Greek gov-
ernment held firm to the very end, despite the fact that more than 90
percent of the Greek population supported the Serbs rather than the
Kosovar Albanians—and held frequent large-scale street demonstra-
tions to show that support. 30 True enough, there remain many un-
knowns about the outlook for NATO’s steadfastness in any future
confrontation along Europe’s eastern periphery. Yet NATO was able
to maintain the one quality that was essential for the success of Allied
______________
29 Interview with Admiral James O. Ellis, USN, commander in chief, Allied Forces,
Southern Europe, Naples, Italy, May 30, 2000. This is not to say, however, that the al-
lies had no intrinsic stake at all. Italy had a stake in preventing further depredations
by Milosevic because of the refugee problem they created. Greece had a major stake
in what happened to the Serbs because of a largely sympathetic population. Germany
also found itself being inundated with refugees. Hungary had good reason to worry
about the Hungarian population still inside Serbia. All of the NATO countries had an
intrinsic interest in stability in Europe, and Milosevic was, if nothing else, a destabi-
lizer of the first order. I am grateful to Alan Gropman of the Industrial College of the
Armed Forces, Washington, D.C., for reminding me of these important facts.
30Air Commodore A. G. B. Vallance, RAF, chief of staff, NATO Reaction Forces (Air)
Staff, Kalkar, Germany, “Did We Really Have a Good War? Myths in the Making,” un-
published manuscript, no date, p. 2.
230 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
31Interview with Air Marshal Sir John Day, RAF, UK Ministry of Defense director of
operations in Allied Force, RAF Innsworth, United Kingdom, July 25, 2000.
32These and other surprises should stand as a sobering reminder that the compara-
tively seamless and unfettered successes achieved by allied air power during Opera-
tion Desert Storm were most likely the exception rather than the rule for future joint
and combined operations—both the operating area and the circumstances surround-
ing the 1991 Gulf War were unique. For more on this point, see Air Vice Marshal Tony
Mason, RAF (Ret.), Air Power: A Centennial Appraisal, London, Brassey’s, 1994, pp.
140–158.
NATO’s Air War in Perspective 231
______________
33Eric Schmitt, “Weak Serb Defense Puzzles NATO,” New York Times, March 26, 1999.
34For a fuller development of this point, see Daniel L. Byman, Matthew C. Waxman,
and Eric Larson, Air Power as a Coercive Instrument , Santa Monica, California, RAND,
MR-1061-AF, 1999. See also Grant T. Hammond, “Myths of the Air War Over Serbia:
Some ‘Lessons’ Not to Learn,” Aerospace Power Journal, Winter 2000, pp. 78–86.
35Cohen and Shelton, “Joint Statement.”
232 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
In fairness to the U.S. and NATO officials most responsible for air op-
erations planning, many of the differences between Allied Force and
the more satisfying Desert Storm experience were beyond the control
of the allies, and they should be duly noted in any critique of the way
the former was conducted. To begin with, as discussed earlier, bad
weather was the rule, not the exception. Second, variegated and
forested terrain limited the effectiveness of many sensors. Third,
Serb SAM operators were more proficient and tactically astute than
those of Iraq. Fourth, alliance complications were greater by far in
Allied Force than were the largely inconsequential intracoalition dif-
ferences during the Gulf War. Finally, because the goal of Allied
Force was more to compel than to destroy, it was naturally more dif-
ficult for senior decisionmakers to measure and assess the air war’s
daily progress, since there was no feedback mechanism to indicate
how well the bombing was advancing toward coercing Milosevic to
comply with NATO’s demands. It was largely for that reason that
most Allied Force planners were surprised when he finally decided to
capitulate.
That said, the most important question with respect to Allied Force
has to do less with platform or systems performance than with the
more basic strategy choices that NATO’s leaders made and what
those choices may suggest about earlier lessons forgotten—not only
from Desert Storm and Deliberate Force but also from Vietnam. Had
Milosevic been content to hunker down and wait out NATO’s
bombing effort, he could easily have challenged the long-term cohe-
sion and staying power of the alliance. Fortunately for the success of
Allied Force, by opting instead to accelerate his ethnic cleansing of
Kosovo, he not only united the West in revulsion but also left NATO
with no alternative but to dig in for the long haul, both to secure an
outcome that would enable the repatriation of nearly a million dis-
placed Kosovars and to ensure its continued credibility as a military
alliance.
Beyond that, the very nature of Operation Allied Force and the man-
ner in which it was conducted from the highest levels both in Wash-
ington and in Brussels placed unique stresses on the JFACC’s ability
to command and control allied air operations. For example, General
Short and his staff had to contend on an unrelenting basis with rapid
shifts in political priorities and SACEUR guidance, as well as with the
myriad pressures occasioned by a random and nonsystematic flow of
assets to the theater, ranging from combat aircraft to staff aug-
mentees in the CAOC. All of these problems emanated from a lack of
consensus among the top decisionmakers on both sides of the At-
lantic as to what the air effort’s military goals were at any given mo-
234 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
ment and what it would take to “prevail.” The de facto “no friendly
loss” rule, stringent collateral damage constraints, and the absence
of a NATO ground threat to force VJ troops to concentrate and thus
make them easier targets further limited the rational employment of
available in-theater assets and placed a premium on accurate infor-
mation and the use of measures that took a disconcertingly long time
to plan, carry out, and evaluate. 36 One realization driven home by
these and other shortcomings was the need for planners in the tar-
geting cell to train together routinely in peacetime before a contin-
gency requires them to react at peak efficiency from the very start.
______________
36 I am indebted to my RAND colleagues James Schneider, Myron Hura, and Gary
McLeod for these on-target summary observations.
NATO’s Air War in Perspective 235
for the success of Allied Force. Not surprisingly, the Serbs were
aware of that fact and were frequently able to exploit it.37
______________
37Cited in Colonel Steve Pitotti, USAF, “Global Environments, Threats, and Military
Strategy (GETM) Update,” Air Armament Summit 2000 briefing, 2000.
38William M. Arkin, “Inside the Air Force, Officers Are Frustrated About the Air War,”
Washington Post, April 25, 1999.
39Rowan Scarborough, “Officers Criticize Air War Strategy,” Washington Times, May
10, 1999.
236 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
40Interview with Air Marshal Sir John Day, RAF, UK Ministry of Defense director of
operations in Allied Force, RAF Innsworth, United Kingdom, July 25, 2000.
41William Drozdiak, “Air War Commander Says Kosovo Victory Near,” Washington
Post, May 24, 1999. Clark himself later indicated that his chief “measure of merit” in
keeping Allied Force on track was “not to lose aircraft, minimize the loss of aircraft.”
He further stated that this exacting desideratum “drove our decisions on tactics, tar-
gets, and which airplanes could participate,” but that it was motivated by a “larger po-
litical rationale: if we wanted to keep this campaign going indefinitely, we had to pro-
tect our air fleet. Nothing would hurt us more with public opinion than headlines that
screamed, ‘NATO LOSES TEN AIRPLANES IN TWO DAYS.’” General Wesley K. Clark,
Waging Modern War: Bosnia, Kosovo, and the Future of Combat, New York, Public
Affairs, 2001, p. 183.
42It bears noting that the zero-loss issue, however seriously it may have been regarded
at the highest leadership levels, had little day-to-day impact on actual combat opera-
tions. As an F-15E instructor Electronic Warfare Officer (EWO) who flew multiple
combat missions with the 494th Fighter Squadron recalled from first-hand experience:
“The issue of the ‘no-losses rule’ did not filter down to the aircrew level, since we al-
ways plan with that goal in mind. We were briefed that there were no ‘high-priority’
targets prior to the opening of hostilities, but that ended up having little effect on the
risk level that we were willing to accept. The concrete effects of the ‘no-loss rule’ were
the 15,000-ft floor and a number of unreasonable ROE restrictions. However, outside
the immediate tactical constraints imposed by the ROE, the prevailing high-level atti-
tude had no effect on tactical operations. We were aware of the priority placed on
NATO’s Air War in Perspective 237
Although the manner in which Allied Force was conducted fell short
of the ideal use of air power, it suggests that gradualism may be here
to stay if U.S. leaders ever again intend to fight wars for marginal or
amorphous interests with as disparate a set of allies as NATO. As the
vice chairman of the JCS at the time, USAF General Joseph Ralston,
noted after the air effort ended, air warfare professionals will con-
tinue to insist, and rightly so, that a massive application of air power
will be more effective than gradualism. Yet, Ralston added, “when
the political and tactical constraints imposed on air use are extensive
and pervasive—and that trend seems more rather than less likely—
then gradualism may be perceived as the only option.”43 General
Jumper likewise intimated that the United States may have little
choice but to accept the burdens of an incremental approach as an
unavoidable cost of working with shaky allies and domestic support
in the future: “It is the politics of the moment that will dictate what
we can do. . . . If the limits of that consensus mean gradualism, then
we’re going to have to find a way to deal with a phased air campaign.
Efficiency may be second.” 44
_____________________________________________________________
minimizing losses, but the effect on the mission was overrated. There were no cases
that I am aware of where the aircrew said, ‘Well, this looks a little hairy, and the prior-
ity is not to lose an airplane, so I won’t do it.’ We were more likely to abort an attack
for collateral damage concerns than we were to abort for survivability issues. As would
be expected, aircrews pressed to the target in the face of serious opposition.” Major
Michael Pietrucha, USAF, personal communication to the author, July 9, 2001.
43“Ralston Sees Potential for More Wars of Gradual Escalation,” Inside the Pentagon,
September 16, 1999, p. 1.
44“Washington Outlook,” Aviation Week and Space Technology, August 23, 1999, p. 27.
It hardly follows, of course, that gradualism and coalitions must invariably be syn-
onymous. They certainly were not in Desert Storm in 1991. Clearly, the extent to
which gradualist strategies will prove unavoidable in the future will depend heavily on
both the shared stakes for would-be coalition partners and the skill of their leaders in
setting the direction and tone of coalition conduct.
238 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
that war is ultimately about politics and that civilian control of the
military is an inherent part of the democratic tradition. It follows
that although airmen and other warfighters are duty-bound to try to
persuade their civilian superiors of the merits of their recommenda-
tions, they also have a duty to live with the hands they are dealt and
to bend every effort to make the most of them in an imperfect
world.45 It also follows that civilian leaders at the highest levels have
an equal obligation to try to stack the deck in such a manner that the
military has the best possible hand to play and the fullest possible
freedom to play it to the best of its ability. This means expending the
energy and political capital needed to develop and enforce a strategy
that maximizes the probability of military success. In Allied Force,
that was not done by the vast majority of the top civilian leaders on
either side of the Atlantic. 46
On the plus side, the air war’s successful outcome despite its many
frustrations suggested that U.S. air power may now have become ca-
pable enough, at least in some circumstances, to underwrite a strat-
egy of incremental escalation irrespective of the latter’s inherent
inefficiencies. What made the gradualism of Allied Force more bear-
able than that of the earlier war in Vietnam is that NATO’s advan-
tages in stealth, precision standoff attack, and electronic warfare
meant that it could fight a one-sided war against Milosevic with
near-impunity and achieve the desired result, even if not in the most
ideal way. 47 That was not an option when U.S. air power was a less
developed tool than it is today.
______________
45On this point, Air Vice Marshal Mason remarked that he had not “spent the past 25
years trying to persuade unbelievers of the efficacy of air power only to finish up
whining because political circumstances made operations difficult.” Personal com-
munication to the author, October 22, 1999. In a similar spirit, the leader of USAFE’s
post–Allied Force munitions effectiveness investigation in Kosovo later suggested that
airmen should “consider a politically restricted target list like the weather: complain
about it, but deal with it.” Colonel Brian McDonald, USAF, briefing at RAND, Santa
Monica, California, December 14, 1999.
46It further follows that airmen, for their part, need to learn not only how to conduct
gradual campaigns more effectively, but also how better to explain convincingly to
politicians the value of using mass and shock early and the greater strategic effective-
ness of effects-based targeting.
47See Colonel Phillip S. Meilinger, USAF, “Gradual Escalation: NATO’s Kosovo Air
Campaign, Though Decried as a Strategy, May Be the Future of War,” Armed Forces
Journal International, October 1999, p. 18.
NATO’s Air War in Perspective 239
On this point, Admiral Ellis, a career fighter pilot himself, was no less
disturbed by the air war’s lethargic pace than was his air component
commander, General Short, or any other airmen on down the line.
However, mindful of the long-standing political and bureaucratic
rule of thumb that “if a problem has no solution, it is no longer a
problem but a fact,” he recognized that ideal-world solutions were
unworkable in the Allied Force setting and that flexibility was re-
quired in applying air doctrine in a difficult situation. As it turned
out, NATO conducted its bombing effort in a way that was not max-
imally efficient, yet that worked in the end to foil Serb strategy, which
was to wait out the alliance and strive mightily to fragment it. Be-
cause the escalation was gradual over time, the coalition succeeded
in holding together. Because NATO used highly conservative tactics,
it lost no aircrews and civilian casualties and collateral damage were
kept to a minimum. In effect, a compromise was struck in which the
air war was intense enough to maintain constant pressure on Milo-
sevic yet measured enough to keep NATO from falling apart. Either
the loss of friendly lives beyond token numbers or an especially grue-
some spectacle of collateral damage could have been more than
enough to incline at least some key allies to call it quits. Noting fur-
ther that NATO fought in this case to establish conditions rather than
to “win” in the classic sense, Ellis added that a campaign strategy
that would have allowed Desert Storm–like intensity and scale of tar-
get attacks to be employed was simply never in the cards.
______________
48Air Commodore A. G. B. Vallance, RAF, “After Kosovo: Implications of Operation Al-
lied Force for Air Power Development,” unpublished paper, p. 4.
49Mueller, “Deus ex Machina?” p. 16.
240 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
On this point, one can fairly suggest that both SACEUR and his
JFACC were equally prone throughout Allied Force to remain wedded
to excessively parochial views of their preferred target priorities,
based on implicit faith in the inherent correctness of their respective
services’ doctrinal teachings. They might more effectively have ap-
proached Milosevic instead as a unique rather than generic oppo-
nent, conducted a serious analysis of his distinctive vulnerabilities,
and then tailored a campaign plan aimed at attacking those vulner-
abilities directly, irrespective of canonical air or land warfare solu-
tions for all seasons. A year after the air war, in a measured reflection
on the recurrent tension that afflicted the interaction of Clark and
Short, Admiral Ellis suggested that the failure of all the services to
advance beyond their propensity to teach only pristine, service-ori-
ented doctrines at their respective war colleges reflected a serious
“cultures” problem and that the services badly need to plan for and
______________
50This includes being held increasingly accountable for their own combat losses. The
Allied Force SEAD experience showed that in crises where less-than-vital U.S. interests
are at stake, near-zero attrition of friendly aircraft and their aircrews will be a high, and
possibly determining, priority governing operational tactics.
NATO’s Air War in Perspective 241
______________
51Conversation with Admiral James Ellis, USN, Headquarters Allied Forces Southern
Europe, Naples, Italy, May 30, 2000.
52William Drozdiak, “Allies Need Upgrade, General Says.” As sensible as this sugges-
tion may have sounded after the fact, however, one must ask how workable it would
have been in actual practice. Wars characteristically feature dynamics that push par-
ticipants beyond anything imaginable at the outset. Setting clear going-in rules is easy
and feasible enough for something short and relatively straightforward, like Operation
Deliberate Force and Operation El Dorado Canyon, the joint USAF-Navy raid on Libya
in 1986. Expecting them in larger and more open-ended operations, however, means
counting on a predictability of events that does not exist in real life. The fact is that
there was a consensus at the start of Allied Force about what was acceptable and what
everyone was willing to do, and that was for 91 targets and two nights of bombing.
NATO’s cardinal error was not its failure to reach a consensus before firing the first
shot; it was its refusal to be honest up front about what it would do if its assumptions
about Milosevic’s resolve proved false. I thank Dr. Daniel Harrington, Office of
History, Hq USAFE, for having shared this insightful observation with me. I would add
that had NATO’s leaders done better at attending to that responsibility, they would
have gone a long way toward satisfying General Short’s expressed concern.
242 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
53Interview with Lieutenant General Michael Short, USAF, PBS Frontline, “War in Eu-
rope,” February 22, 2000.
54Bradley Graham, “General Says U.S. Readiness Is Ailing,” Washington Post, April 30,
1999.
55General John Jumper, USAF, “Oral Histories Accomplished in Conjunction with Op-
eration Allied Force/Noble Anvil.”
56Drozdiak, “Allies Need Upgrade, General Says.”
NATO’s Air War in Perspective 243
that the latter had begun to operate with some success in the end
“made the Yugoslav army come out and fight and try to blunt their
offensive. . . . And once they moved, or fired their artillery, our
strikers learned where they were and could go in for the kill.”57 Had
VJ forces in Kosovo faced an imminent NATO ground invasion, or
even a credible threat of such an invasion later, they would have
been obliged to move troops and supplies over bridges that NATO
aircraft could have dropped. They also would have been compelled
to concentrate and maneuver in ways that would have made it easier
for NATO to find and attack them.
______________
57Ibid.
58Doyle McManus, “Clinton’s Massive Ground Invasion That Almost Was,” Los Ange-
les Times, June 9, 2000.
244 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
need actually to commit NATO troops to battle in the end. The mere
fact of a serious Desert Shield–like deployment of NATO ground
troops along the Albanian and Macedonian borders would have
made their VJ counterparts more easily targetable by allied air power.
Had such a deployment commenced in earnest, it also might have
helped to deter, or at least lessen, the ethnic cleansing of Kosovo by
giving VJ troops a more serious concern to worry about. In both
cases, it could have enabled a quicker end to the war.
Indeed, well before Allied Force ended, there was a gathering sense
among some observers that Serbia’s ground forces were being given
more credit than they deserved as an excuse for ruling out a NATO
land-invasion option. As one former U.S. Army officer pointed out,
Milosevic’s army was a small conscript-based force with an active
component of only some 115,000 troops who relied on antiquated
Soviet equipment, mainly the 1950s-vintage T-55 tank. Air strikes
during the first few nights of Allied Force had already rendered Yu-
goslavia’s small air force a non-factor in any potential NATO ground
push. The VJ’s petroleum and other stocks for sustainment had also
been rapidly depleted by the bombing, leaving the Serbs with, at
best, only a minimal capacity to wage conventional war against a se-
rious ground opponent. In contrast, the modern and well-equipped
NATO ground forces arguably possessed enough combat power “to
make mincemeat of the Yugoslav army.”59
______________
59Andrew J. Bacevich, “Target Belgrade: Why a Ground War Would Be a Rout,” Na-
tional Review, May 3, 1999, p. 29.
NATO’s Air War in Perspective 245
______________
60General Merrill A. McPeak, USAF (Ret.), “The Kosovo Result: The Facts Speak for
Themselves,” Armed Forces Journal International, September 1999, p. 64. In a similar
vein, the chief of staff of the RAF later faulted NATO’s decision to rule out a ground
option from the start of the air war as “a strategic mistake” that enabled Serb forces to
forgo preparing defensive positions, hide their tanks and artillery and make maximum
use of deception against NATO attack efforts, and conduct their ethnic cleansing of
Kosovo with impunity. Michael Evans, “Ground War ‘Error,’” London Times, March
24, 2000.
246 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
war’s success.61 Ellis charged that NATO’s leaders “called this one
absolutely wrong” and that their failure to anticipate what might oc-
cur once their initial strategy of hope failed occasioned most of the
untoward consequences that ensued thereafter.62 These included
the hasty activation of a joint task force, a race to find suitable tar-
gets, an absence of coherent campaign planning, and lost opportu-
nities caused by the failure to think through unpleasant excursions
from what had been expected. Ellis concluded that the imperatives
of consensus politics within NATO made for an “incremental war”
rather than for “decisive operations,” that excessive concern over
collateral damage created “sanctuaries and opportunities for the
adversary—which were successfully exploited,” and that the lack of a
credible NATO ground threat “probably prolonged the air cam-
paign.”63 It was only because Milosevic made a blunder no less tow-
ering than NATO’s preclusion of a ground option that the war had
the largely positive outcome that it did.
Indeed, that NATO prevailed in the end with only two aircraft lost
and no combat fatalities sustained surely reflected good fortune at
least as much as the professionalism of its aircrews and their com-
manders. General Jumper explained afterward that “we set the bar
fairly high when we fly more than 30,000 combat sorties and we don’t
lose one pilot. It makes it look as if air power is indeed risk free and
too easy a choice to make.” Amplifying on the same point, retired
RAF Air Vice Marshal Tony Mason observed that seeking to minimize
one’s losses is both admirable and proper up to a point, yet it can
lead to self-deterrence when efforts to escape the costs of war are
______________
61Amplifying on his suggestion that luck was the key player, Ellis pointed out how
much worse matters would have been for the alliance had NATO experienced any one
of a number of untoward developments: an enemy attack on its troops deployed in
theater with ground forces or tactical ballistic missiles; the possibility of even a few
NATO aircrews being killed in action or captured as POWs; the continuation of the
fighting into the winter; the depletion of U.S. precision munitions stocks; the weaken-
ing or evaporation of public support; an allied ground invasion becoming the only op-
tion; or a decision by France or Italy to withdraw from further participation.
62Revealingly, barely a week into Allied Force, one senior Clinton administration offi-
cial, when asked what NATO’s strategy would be should Phase III of the air war fail to
persuade Milosevic to admit defeat, replied: “There is no Phase IV.” Quoted in John
Broder, “In Grim Week, Pep Talk from the President,” New York Times, April 1, 1999.
63Elaine M. Grossman, “For U.S. Commander in Kosovo, Luck Played Role in Wartime
Success,” Inside the Pentagon, September 9, 1999, p. 1.
NATO’s Air War in Perspective 247
______________
64Comments at an Air Force Association Eaker Institute colloquy, “Operation Allied
Force: Strategy, Execution, Implications,” held at the Ronald Reagan International
Trade Center, Washington, D.C., August 16, 1999.
65These were among numerous other conclusions suggested by Major General Ronald
Keys, USAF, director of operations (J-3), U.S. European Command, cited in Colonel
Steve Pitotti, USAF, “Global Environments, Threats, and Military Strategy (GETM) Up-
date,” Air Armament Summit 2000 briefing, Eglin AFB, Florida, 2000.
248 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
As for the ultimate wisdom of the allied decision to proceed with the
air war in the first place, the United States and NATO displayed
an ability in this case to apply coercion successfully through air
power from a poorly prepared battlefield at a remarkably low cost in
noncombatant fatalities caused by direct collateral damage.68 Yet
there is a danger that making a habit of such displays by accepting
Allied Force as a model for future interventions could easily lead to
an erosion of the U.S. claim to global leadership.69 On the contrary,
Allied Force should have underscored the fact that one of the most
acute challenges facing U.S. policymakers in the age of a single su-
perpower entails deciding when, and in what manner, to intervene in
humanitarian crises that do not yet impinge directly on U.S. security
interests.
______________
66Work on this is being performed by Alan Vick of RAND.
67Elaine M. Grossman, “U.S. Military Debates Link Between Kosovo Air War, Stated
Objectives,” Inside the Pentagon, April 20, 2000, p. 6.
68A heated argument arose after the war ended between defenders and critics of the
Clinton administration’s strategy for Kosovo over whether the approach taken, despite
its low cost in noncombatant lives lost to direct collateral damage, nonetheless
produced an unconscionably high loss of civilian innocents to the Serbian ethnic
cleansing campaign which it allegedly accelerated. For a snapshot summary of the
positions taken on both sides, see Christopher Layne and Benjamin Schwarz, “Kosovo
II: For the Record,” The National Interest, Fall 1999, pp. 9–15, and Ivo Daalder, “NATO
and Kosovo,” The National Interest, Winter 1999/2000, pp. 113–117.
69I am grateful to Lieutenant General Bradley Hosmer, USAF (Ret.), for bringing this
point to my attention.
NATO’s Air War in Perspective 249
that air power prevailed despite a U.S. leadership that was unwilling
to take major risks and an alliance that held together only with often
paralyzing drag. Fortunately, the Clinton administration did a cred-
itable job of keeping the allies together in the end, albeit at the cost of
what Brent Scowcroft called “a bad strategy” that raised basic ques-
tions about the limits of alliance warfare and about whether the
United States should, in the future, settle instead for coalitions of the
willing, at least in less than the cataclysmic showdowns of the sort
that NATO was initially created to handle.70 One can only wonder
what greater efficiencies might have been registered by a more as-
sertive campaign approach had the U.S. government been willing to
play a more proactive role in leading from the front and setting both
the direction and pace for NATO’s more hesitant allies.71
Lesson One from both Vietnam and Desert Storm should have been
that one must not commit air power in “penny packets,” as the
British say, to play less-than-determined games with the risk calculus
of the other side. Although it can be surgically precise when preci-
sion is called for, air power is, at bottom, a blunt instrument
designed to break things and kill people in pursuit of clear and
militarily achievable objectives. Not without reason have air warfare
professionals repeatedly insisted since Vietnam that if all one wishes
to do is to “send a message,” call Western Union. On this point, Eliot
Cohen summed it up well five years before the Kosovo crisis erupted
when he compared air power’s lately acquired seductiveness to
modern teenage romance in its seeming propensity to offer political
leaders a sense of “gratification without commitment.”72
______________
70John F. Harris, “Berger’s Caution Has Shaped Role of U.S. in War,” Washington Post,
May 16, 1999.
71In a measured indictment of the Clinton administration’s comportment in this re-
gard, two Brookings Institution analysts wrote that “what was missing . . . was less al-
lied will than a demonstrated American ability and willingness to lead a joint effort.
NATO works best when Washington knows what it wants done and leads the effort to
get the alliance there. In the runup to the Kosovo war, both elements were tragically
lacking. . . . Although it is impossible to know whether the allies would have gone
along with a more robust strategy, including early use of ground forces, the United
States never made the case. U.S. policy presumed the allies’ rejection, just as it
presumed congressional opposition to the use of ground forces.” Ivo H. Daalder and
Michael E. O’Hanlon, Winning Ugly: NATO’s War to Save Kosovo, Washington, D.C.,
Brookings Institution, 2000, pp. 98, 222.
72Eliot A. Cohen, “The Mystique of U.S. Air Power,” Foreign Affairs, January/February
1994, p. 109.
250 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
______________
73Joseph Fitchett, “Clark Recalls ‘Lessons’ of Kosovo,” International Herald Tribune,
May 3, 2000.
BIBLIOGRAPHY
OFFICIAL PUBLICATIONS
Air Force Basic Doctrine, Maxwell AFB, Alabama, Headquarters Air
Force Doctrine Center, AFDD-1, September 1997.
The Air War Over Serbia: Aerospace Power in Operation Allied Force,
Washington, D.C., Headquarters United States Air Force, April 1,
2000.
CONGRESSIONAL TESTIMONY
Clark, General Wesley, USA, testimony to the Senate Armed Services
Committee, Washington, D.C., July 1, 1999.
251
252 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
BOOKS
Andric, Ivo, The Bridge on the Drina, Chicago, University of Chicago
Press, 1977.
Clark, General Wesley K., Waging Modern War: Bosnia, Kosovo, and
the Future of Combat, New York, Public Affairs, 2001.
Glenny, Misha, The Balkans: Nationalism, War and the Great Powers,
1809–1999, New York: Penguin Books, 2000.
Ignatieff, Michael, Virtual War: Kosovo and Beyond, New York, Henry
Holt and Company, Inc., 2000.
Judah, Tim, Kosovo: War and Revenge, New Haven, Connecticut, Yale
University Press, 2000.
Mason, Air Vice Marshal Tony, RAF (Ret.), Air Power: A Centennial
Appraisal, London, Brassey’s, 1994.
Rackham, Peter, ed., Jane’s C4I Systems, 1994–95, London, Jane’s In-
formation Group, 1994.
Thompson, Wayne, To Hanoi and Back: The U.S. Air Force and North
Vietnam, 1966–1973, Washington, D.C., Smithsonian Institute
Press, 2000.
Peters, John E., Stuart Johnson, Nora Bensahel, Timothy Liston, and
Traci Williams, European Contributions to Operation Allied Force:
Implications for Transatlantic Cooperation, Santa Monica, Cali-
fornia, RAND, MR-1391-AF, 2001.
Stephens, Alan, Kosovo, or the Future of War, Paper Number 77, Air
Power Studies Center, RAAF Fairbairn, Australia, August 1999.
254 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
White, Lieutenant Colonel Paul K., USAF, Crises After the Storm: An
Appraisal of U.S. Air Operations in Iraq Since the Persian Gulf War,
Military Research Papers No. 2, Washington, D.C., Washington In-
stitute for Near East Policy, 1999.
________, “Top Air Force Leaders to Get Briefed on Serbia Air War
Report,” Defense Daily, June 13, 2000.
Aubin, Stephen P., “Newsweek and the 14 Tanks,” Air Force Maga-
zine, July 2000.
________, “MiGs Tried to Shoot Down Air Force Tanker over Bosnia,”
Defense Week, May 17, 1999.
Cohen, Eliot A., “The Mystique of U.S. Air Power,” Foreign Affairs,
January/February 1994.
Crawley, Vince, “Air Force Needs 90 Days Between Wars, Chief Says,”
Defense Week, August 9, 1999.
Foote, Sheila, “Shelton: Risk Was the Key in Decision Not to Use
Apaches,” Defense Daily, September 10, 1999.
Hammond, Grant T., “Myths of the Air War Over Serbia: Some
‘Lessons’ Not to Learn,” Aerospace Power Journal, Winter 2000.
“Hope for the Best, and a Spot of Golf,” The Economist, April 3, 1999.
McPeak, General Merrill A., USAF (Ret.), “The Kosovo Result: The
Facts Speak for Themselves,” Armed Forces Journal International,
September 1999.
________, and Robert Wall, “NATO Vows Air Strikes Will Go the
Distance,” Aviation Week and Space Technology, March 29, 1999.
Murphy, Vice Admiral Daniel J., USN, “The Navy in the Balkans,” Air
Force Magazine, December 1999.
260 NATO’s Air War for Kosovo: A Strategic and Operational Assessment
Newman, Richard J., “The Bombs That Failed in Kosovo,” U.S. News
and World Report, September 20, 1999.
________, “In the Skies over Serbia,” U.S. News and World Report,
May 24, 1999.
________, “U.S. Troops Edge Closer to Kosovo,” U.S. News and World
Report, June 7, 1999.
Ripley, Tim, “Harriers over the Kosovo ‘Kill Boxes,’” World Air Power
Journal, Winter 1999/2000.
Scott, William B., “JTIDS Provides F-15Cs ‘God’s Eye View,’” Aviation
Week and Space Technology, April 29, 1996.
Seigle, Greg, “Prowler Jammers Used to Aid NATO Air Assault,” Jane’s
Defense Weekly, March 31, 1999.
Sharer, Commander Wayne D., USN, “The Navy’s War over Kosovo,”
Proceedings, U.S. Naval Institute, October 1999.
Stroup, Lieutenant General Theodore G., Jr., USA (Ret.), “Task Force
Hawk: Beyond Expectations,” Army Magazine, August 1999.
Tirpak, John A., “The First Six Weeks,” Air Force Magazine, June 1999.
“U.S. Mobilizes Guard, Reserve for Balkan Duty,” Air Force Magazine,
June 1999.
“Verbatim Special: The Balkan War,” Air Force Magazine, June 1999.
NEWSPAPER ARTICLES
Andrews, Edmund L., “Aboard Advanced Radar Flight, U.S. Watches
Combat Zone,” New York Times, June 14, 1999.
Apple, R. W., Jr., “With Decision to Attack, a New Set of U.S. Goals,”
New York Times, March 25, 1999.
________, “Inside the Air Force, Officers Are Frustrated About the Air
War,” Washington Post, April 25, 1999.
________, “They’re Unmanned, They Fly Low, and They Get the Pic-
ture,” New York Times, June 3, 1999.
Block, Robert, “In Belgrade, Hardship Grows Under Sustained Air As-
sault,” Wall Street Journal, May 12, 1999.
________, “How to Lay Doubt Aside and Put the Best Face on a Bad
Week in the Balkans,” New York Times, April 1, 1999.
________, “In Grim Week, Pep Talk from the President,” New York
Times, April 1, 1999.
Carroll, James, “The Truth About NATO’s Air War,” Boston Globe,
June 20, 2000.
Cox, James, “Poll: Mission Isn’t Seen as U.S. Victory,” USA Today,
June 15, 1999.
Deans, Bob, “Pentagon Mum About Air Mission,” European Stars and
Stripes, April 27, 1999.
________, “Over Balkans, It’s Beauty vs. the Beast,” Norfolk Virginian-
Pilot, April 26, 1999.
Gellman, Barton, “Key Sites Pounded for 2nd Day,” Washington Post,
March 26, 1999.
Gertz, Bill, “Remote Radar Allows Serbs to Keep Firing at NATO Jets,”
Washington Times, April 13, 1999.
Bibliography 267
Gordon, Michael R., “Allied Air Chief Stresses Hitting Belgrade Sites,”
New York Times, May 13, 1999.
________, “NATO to Hit Serbs from 2 More Sides,” New York Times,
May 11, 1999.
________, “A War out of the Night Sky: 10 Hours with a Battle Team,”
New York Times, June 3, 1999.
________, and Eric Schmitt, “Shift in Targets Let NATO Jets Tip the
Balance,” New York Times, June 5, 1999.
________, and Dana Priest, “‘No Way to Fight a War’: Hard Lessons of
Air Power, Coalitions,” Washington Post, June 6, 1999.
Harris, John F., “Berger’s Caution Has Shaped Role of U.S. in War,”
Washington Post, May 16, 1999.
Hedges, Chris, “Angry Serbs Hear a New Explanation: It’s All Russia’s
Fault,” New York Times, July 16, 1999.
Kaminski, Matthew, and John Reed, “NATO Link to KLA Rebels May
Have Helped Seal Victory,” Wall Street Journal, July 6, 1999.
King, Neil, Jr., “War Against Yugoslavia Lapses into Routine, but
Clock Is Ticking,” Wall Street Journal, May 6, 1999.
Krepinevich, Andrew, “Two Cheers for Air Power,” Wall Street Jour-
nal, June 11, 1999.
Lippman, Thomas W., and Dana Priest, “NATO Builds Firepower for
24-Hour Attacks,” Washington Post, March 30, 1999.
Loeb, Vernon, and Steven Mufson, “CIA Analyst Raised Alert on Chi-
na’s Embassy,” Washington Post, June 24, 1999.
Moore, Molly, and Bradley Graham, “NATO Plans for Peace, Not
Ground Invasion,” Washington Post, May 17, 1999.
Peltz, James, and Jeff Leeds, “Stealth Fighter’s Crash Reveals a De-
sign’s Limits,” Los Angeles Times, March 30, 1999.
Pfaff, William, “After NATO’s Lies About Kosovo, It’s Time to Come
Clean,” International Herald Tribune, May 11, 2000.
Rice, Donald B., “No Stealth to Pentagon’s Bias Against the B-2,” Los
Angeles Times, May 9, 1999.
________, “B-2 Drops Its Bad PR in Air War,” Los Angeles Times, July 8,
1999.
________, “Milosevic War Machine Has a Lot of Fight Left,” Los Ange-
les Times, April 29, 1999.
________, “Time Is Not on the Side of U.S., Allies,” Los Angeles Times,
March 25, 1999.
________, “U.S. Pilots Face Perilous Task, Pentagon Says,” Los Angeles
Times, March 20, 1999.
________, “Use of Ground Troops Not Fully Ruled Out,” Los Angeles
Times, March 29, 1999.
Schmitt, Eric, “Aim, Not Arms, at the Root of Mistaken Strike on Em-
bassy,” New York Times, May 10, 1999.
________, “New Army Chief Seeks More Agility and Power,” New York
Times, June 24, 1999.
________, and Steven Lee Myers, “NATO Said to Focus Raids on Serb
Elite’s Property,” New York Times, April 19, 1999.
________, and Molly Moore, “Plan for Kosovo Pullout Signed,” Wash-
ington Post, June 10, 1999.
Suro, Roberto, and Thomas E. Ricks, “Pentagon: Kosovo Air War Data
Leaked,” Washington Post, March 10, 2000.
Whitney, Craig R., “NATO Chief Admits Bombs Fail to Stem Serb
Operations,” New York Times, April 28, 1999.
________, “NATO Had Signs Its Strategy Would Fail Kosovars,” New
York Times, April 1, 1999.
________, “On 7th Day, Serb Resilience Gives NATO Leaders Pause,”
New York Times, March 31, 1999.
Wilson, George C., “Memo Says Apaches, Pilots Were Not Ready,”
European Stars and Stripes, June 20, 1999.
BRIEFINGS
Baldazzi, Colonel E., Italian Air Force, “Host Nation Support for the
Kosovo Air Campaign,” briefing at a conference on “The NATO
Joint Force Air Component Commander Concept in Light of the
Kosovo Air Campaign,” Headquarters NATO Reaction Force Air
Staff, Kalkar, Germany, December 1–3, 1999.
Ellis, Admiral James O., USN, commander in chief, U.S. Naval Forces,
Europe, and commander, Allied Forces Southern Europe and Joint
Task Force Noble Anvil, “The View from the Top,” 1999.
Watts, Barry D., “The EA-6B, E-8C, and B-2 in Operation Allied
Force,” Northrop Grumman Analysis Center, Rosslyn, Virginia,
May 8, 2000.